Guidelines on Clinical Evaluation of Vaccines: Regulatory ... · 7 Guidelines on Clinical...

1

2

WHO/DRAFT/27 January 2016 3

ENGLISH ONLY 4

5

6

Guidelines on Clinical Evaluation of Vaccines: Regulatory Expectations 7

Proposed revision of WHO TRS 924, Annex 1 8

9

NOTE: 10

11 This document has been prepared for the purpose of inviting comments and suggestions on the proposals 12 contained therein, which will then be considered by the Expert Committee on Biological Standardization 13 (ECBS). Publication of this early draft is to provide information about the proposed Guidelines on 14 Clinical Evaluation of Vaccines: Regulatory Expectations, to a broad audience and to improve 15 transparency of the consultation process. 16 17 The text in its present form does not necessarily represent an agreed formulation of the Expert 18 Committee. Written comments proposing modifications to this text MUST be received by 15

th 19

March 2016 in the Comment Form available separately and should be addressed to the World Health 20 Organization, 1211 Geneva 27, Switzerland, attention: Department of Essential Medicines and Health 21 Products (EMP). Comments may also be submitted electronically to the Responsible Officer: Dr Ivana 22 Knezevic at email: [email protected]. 23 24 The outcome of the deliberations of the Expert Committee on Biological Standardization will be 25 published in the WHO Technical Report Series. The final agreed formulation of the document will be 26 edited to be in conformity with the "WHO style guide" (WHO/IMD/PUB/04.1). 27

28

© World Health Organization 2016 29

All rights reserved. Publications of the World Health Organization can be obtained from WHO Press, World Health 30 Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791 4857; e-mail: 31 [email protected]). Requests for permission to reproduce or translate WHO publications – whether for sale or for non-32 commercial distribution – should be addressed to WHO Press, at the above address (fax: +41 22 791 4806; e-mail: 33 [email protected]). 34

The designations employed and the presentation of the material in this publication do not imply the expression of any 35 opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or 36

mailto:[email protected]

mailto:[email protected]

WHO/DRAFT/27 January 2016

area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent 37 approximate border lines for which there may not yet be full agreement. 38 39 The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or 40 recommended by the World Health Organization in preference to others of a similar nature that are not mentioned. Errors 41 and omissions excepted, the names of proprietary products are distinguished by initial capital letters. 42 43 All reasonable precautions have been taken by the World Health Organization to verify the information contained in this 44 publication. However, the published material is being distributed without warranty of any kind, either expressed or implied. 45 The responsibility for the interpretation and use of the material lies with the reader. In no event shall the World Health 46 Organization be liable for damages arising from its use. 47

48 The named authors [or editors as appropriate] alone are responsible for the views expressed in this publication. 49

50

51

52

53

54

Recommendations and guidelines published by WHO are intended to be scientific and advisory in

nature. Each of the following sections constitutes guidance for national regulatory authorities (NRAs)

and for manufacturers of biological products. If an NRA so desires, these Guidelines may be adopted

as definitive national requirements, or modifications may be justified and made by the NRA. It is

recommended that modifications to these Guidelines be made only on condition that modifications

ensure that the vaccine is at least as safe and efficacious as that prepared in accordance with the

recommendations set out below. The parts of each section printed in small type are comments or

examples for additional guidance intended for manufacturers and NRAs, which may benefit from

those details.


Table of Contents 55

56

1. Introduction 57

2. Scope 58

3. Glossary 59

4. Vaccine Clinical Development Programs 60

4.1 General considerations 61

4.1.1 Consultation with National Regulatory Authorities (NRAs) 62

4.1.2 Use of independent data monitoring committees 63

4.1.3 Registering and reporting clinical trials 64

4.2 New candidate vaccines 65

4.2.1 Safety and immunogenicity trials 66

4.2.1.1 Initial trials 67

4.2.1.2 Further trials 68

4.2.1.3 Confirmatory (or pivotal) trials 69

4.2.2 Efficacy trials 70

4.2.3 Pivotal safety trials 71

4.3 Post-licensure clinical evaluations 72

5. Immunogenicity 73


5.2 Characterization of the immune response 75

5.3 Measuring the immune response 76

5.2.1 Collection of specimens 77

5.2.2 Immunological parameters 78

5.2.2.1 Humoral immune response 79

5.2.2.2 Cell-mediated immune response 80

5.2.3 Assays 81

5.4 Determination and use of immunological correlates of protection 82

5.4.1 Immunological correlates of protection and their uses 83

5.4.2 Establishing an ICP 84

5.5 Immunogenicity trials 85


5.5.1 Objectives 86

5.5.2 General considerations for trial designs 87

5.5.2.1 Endpoints 88

5.5.2.2 Exploratory trials 89

5.5.2.3 Superiority trials 90

5.5.2.4 Non-inferiority trials 91

5.5.3 Analysis and interpretation 92

5.6 Specific considerations for trial design and interpretation 93

5.6.1 Selection of formulation and posology 94

5.6.1.1 Selecting the formulation and posology for initial licensure 95

5.6.1.2 Amending or adding posologies after initial licensure 96

5.6.1.3 Post-primary doses 97

5.6.2 Using immunogenicity data to predict efficacy 98

5.6.2.1 Bridging to efficacy data 99

5.6.2.2 Other approaches 100

5.6.3 Co-administration trials 101

5.6.4 Immunization of pregnant women 102

5.6.4.2 Dose-finding in pregnancy 103

5.6.4.1 Aims of immunization during pregnancy 104

5.6.4.3 Passive protection of infants 105

5.6.5 Changes to the manufacturing process 106

5.6.6 Lot-to-lot consistency trials 107

6. Efficacy and effectiveness 108

6.1 Approaches to determination of efficacy 109

6.1.1 Human challenge trials 110

6.1.2 Preliminary efficacy trials 111

6.1.3 Confirmatory (pivotal) efficacy trials 112

6.2 Design and conduct of efficacy trials 113

6.2.1 Selection of trial sites 114

6.2.2 Candidate (test) vaccine group(s) 115

6.2.3 Control (reference) group(s) 116


6.2.3.1 Control groups not vaccinated against the infectious disease to be 117

prevented 118

6.2.3.2 Control groups vaccinated against the infectious disease to be 119

prevented 120

6.2.4 Trial designs 121

6.2.4.1 Randomization 122

6.2.4.2 Types of trial design 123

6.2.5 Clinical endpoints 124

6.2.5.1 Primary endpoints 125

6.2.5.2 Secondary endpoints 126

6.2.6 Case definition 127

6.2.7 Case ascertainment 128

6.2.8 Statistical considerations 129

6.2.8.1 Sample size 130

6.2.8.2 Analysis populations 131

6.2.8.3 Primary analysis 132

6.2.8.4 Other analyses 133

6.2.8.5 Other issues 134

6.3 Approaches to determination of effectiveness 135

7. Safety 136


7.2 Assessment of safety in clinical trials 138

7.2.1 Safety as a primary or secondary endpoint 139

7.2.1.1 Safety as a primary endpoint 140

7.2.1.2 Safety as a secondary endpoint 141

7.2.2 Recording and reporting adverse events 142

7.2.2.1 Methods 143

7.2.2.2 Solicited signs and symptoms 144

7.2.2.3 Unsolicited AEFIs 145

7.2.2.4 Other investigations 146

7.2.3 Categorization of adverse events 147


7.2.3.1 Causality 148

7.2.3.2 Severity 149

7.2.3.3 Other categorisation 150

7.2.4 AE reporting rates within and between trials 151

7.3 Size of the pre-licensure safety database 152

7.4 Post-licensure safety surveillance 153

154

Authors and Acknowledgements 155

References 156

Appendix 1. Human challenge trials 157

158


1. Introduction 159

160

This guideline is intended to replace WHO Technical Report, Series No. 924, Annex 1 Guidelines 161

on clinical evaluation of vaccines: Regulatory Expectations, which was adopted by the Expert 162

Committee on Biological Standardization (ECBS) in 2001 (1). This document of 2001 has 163

served as a basis for setting or updating national requirements for the evaluation and licensing of 164

a broad range of vaccines as well as for WHO vaccine prequalification. 165

166

Following on the establishment of the document of 2001, more than 20 vaccine-specific 167

documents that include a section on clinical evaluation have been adopted by the ECBS, all of 168

which are intended to be read in conjunction with TRS 924, Annex 1 (2). These include 169

documents that address polio vaccines [OPV, IPV], whole cell pertussis and acellular pertussis 170

vaccines, meningococcal conjugate vaccines for serotypes A and C, pneumococcal conjugate 171

vaccines and vaccines intended to prevent diseases due to rotaviruses, dengue viruses, human 172

papillomaviruses and malaria parasites. 173

174

This guideline has been prepared to reflect the scientific and regulatory experience that has been 175

gained from vaccine clinical development programs since the adoption of the above mentioned 176

version in 2001. Many challenging issues surrounding appropriate and feasible vaccine clinical 177

development programs for specific types of vaccines have arisen in the intervening period. For 178

example, there has been increasing recognition of the potential need to base initial licensure of 179

certain vaccines on safety and immunogenicity data only (i.e. it is not feasible to generate pre-180

licensure efficacy data) and in the absence of an established immunological correlate of 181

protection (ICP). 182

183

This guideline is intended for use by national regulatory authorities (NRAs), companies 184

developing and holding licences for vaccines, clinical researchers and investigators. It considers 185

the variable content of clinical development programs, clinical trial designs, the interpretation of 186

trial results and post-licensing activities. The content of the various sections is intended to assist 187

in the preparation and approval of clinical trial applications, applications for initial licensure and 188


applications to support post-licensure changes as well as to provide guidance on post-licensure 189

activities, such as pharmacovigilance and estimation of vaccine effectiveness. 190

191

The main changes (modification or expansion of previous text and additional issues covered) in 192

this revision compared to the above mentioned version of TRS No. 924, Annex 1, 2001 (1) 193

include, but are not limited to, the following: 194

195

Immunogenicity 196

General principles for comparative immunogenicity studies, including selection of the 197

comparators, endpoints and acceptance criteria for concluding non-inferiority or 198

superiority of immune responses 199

Situations in which age de-escalation studies may be inappropriate 200

Assessing the need for and timing of post-primary doses 201

Using different vaccines for priming and boosting 202

Assessing the ability of vaccines to elicit immune memory or to cause hypo-203

responsiveness 204

Using immunogenicity data to predict vaccine efficacy, with or without bridging to 205

efficacy data 206

The derivation and uses of immunological ICPs 207

Vaccination of pregnant women to protect them and/or their infants 208

209

Efficacy 210

Role and potential value of human challenge studies 211

Need for and feasibility of conducting vaccine efficacy studies 212

Selection of appropriate control groups in different circumstances 213

Comparing extended with parent versions of vaccines 214

Predicting vaccine efficacy when there is no ICP and vaccine efficacy studies are not 215

feasible 216

Preliminary and confirmatory vaccine efficacy studies and their design 217

Vaccines with modest efficacy and/or that provide a short duration of protection 218


Extrapolating data between geographic/genetically diverse populations 219

Role of sponsors and public health authorities in generating vaccine effectiveness data 220

221

Safety 222

Detailed consideration of the collection and analysis of safety data from clinical trials 223

Consideration of size of the pre-licensure database by type of vaccine and its novelty 224

Consideration of the safety database by population sub-group 225

Special safety considerations by vaccine construct 226

Circumstances of limited safety data pre-licensure 227

Use of vaccine registries and disease registries 228

Particular issues for vaccine pharmacovigilance activities 229

230

Due to the fact that a separate document on nonclinical evaluation of vaccines was established 231

in 2003 (3), the section on that topic in the 2001 version has been removed. Furthermore, the 232

structure of the document has changed. In particular, a number of methodological 233

considerations have now been incorporated into relevant sections and subsections rather than 234

being described in a separate section. In line with the changes made in the document, the 235

Glossary and References have been updated. 236

237

The WHO has also made available several other guidelines of relevance to clinical development 238

programs for vaccines. These should be consulted as appropriate and include: 239

Good clinical practice for trials on pharmaceutical products (4) 240

Good manufacturing practice for pharmaceutical preparations (5) 241

Good manufacturing practice for biological products (6) 242

Guidelines on nonclinical evaluation of vaccines (3) 243

Guidelines on nonclinical evaluation of vaccine adjuvants and adjuvanted vaccines (7) 244

Guidelines on procedures and data requirements for changes to approved vaccines (8) 245

Guidelines for independent lot release of vaccines by regulatory authorities (9) 246

Recommendations for the evaluation of animal cell cultures as substrates for the 247

manufacture of biological medicinal products and for the characterization of cell banks (10) 248


Clinical Considerations for Evaluation of Vaccines for Prequalification (11) 249

The WHO manual Immunization in practice (12) 250

WHO expert consultation on the use of placebos in vaccine trials (13) 251

252

Furthermore, guidance on various aspects of pre-licensure clinical development programs for 253

vaccines and post-licensure assessment is also available from several other bodies, such as the 254

International Conference on Harmonization (ICH), the European Medicines Agency (EMA), the 255

United States Food and Drug Administration (FDA) and the United Kingdom Medical Research 256

Council (MRC). These WHO guidelines are not intended to conflict with, but rather to 257

complement, these other documents. 258

259

2. Scope 260

261

This guideline considers clinical development programmes for vaccines that are intended to 262

prevent infectious diseases in humans by eliciting protective immune responses that are 263

sufficient to prevent clinically apparent infections. It includes vaccines that may be given before 264

exposure or shortly after known or presumed exposure to an infectious agent to prevent onset of 265

clinical disease. Protective immune responses may be directed against one or more specific 266

antigenic components of micro-organisms or against substances produced and secreted by them 267

(e.g. toxins) that are responsible for clinical disease. 268

269

The guideline is applicable to vaccines which contain one of more of the following: 270

Microorganisms that have been inactivated by chemical and/or physical means 271

Live microorganisms that have been rendered avirulent in humans as a result of attenuation 272

processes or specific genetic modification 273

Antigenic substances that have been derived from micro-organisms. These may be purified 274

from micro-organisms and used in their natural state or may be modified (e.g. detoxified by 275

chemical or physical means, aggregated or polymerized). 276

Antigens that have been manufactured by synthetic processes or produced by live organisms 277

using recombinant DNA technology. 278


Antigens (however manufactured) that have been chemically conjugated to a carrier 279

molecule to modify the interaction of the antigen with the host immune system. 280

Antigens that are expressed by another micro-organism which itself does not cause clinical 281

disease but acts as a live vector (e.g. live viral vectored vaccines, live attenuated chimeric 282

vaccines). 283

In addition, although naked DNA vaccines are not specifically discussed in this guideline the 284

principles and development programs outlined are broadly applicable. 285

286

This guideline does not apply to: 287

Therapeutic vaccines (i.e. used for treatment of disease) 288

Vaccines intended for any purpose other than prevention of infectious diseases and the 289

consequences of infectious diseases. 290

291

3. Glossary 292

293

The definitions given below apply to the terms used in this guideline. They may have different 294

meanings in other contexts. 295

296

Adverse event (AE) 297

Any untoward medical occurrence in a trial subject. An AE does not necessarily have a causal 298

relationship with the vaccine. 299

300

Adverse event following immunization (AEFI) 301

Any untoward medical occurrence that follows immunization using a licensed vaccine outside of 302

a clinical trial setting. An AEFI does not necessarily have a causal relationship with the use of 303

the vaccine. The AEFI may be any unfavourable or unintended sign, abnormal laboratory 304

finding, symptom or disease. 305

306

Attack rate 307

The proportion of the population exposed to an infectious agent who become (clinically) ill. 308

309


Blinding 310

A procedure in which one or more parties involved in a clinical trial are kept unaware of the 311

treatment assignment(s). Double blinding refers to the vaccinees/care-givers, investigator(s) and 312

sponsor staff being unaware of the treatment assignment during the conduct of the trial and at 313

least until after completion of the primary analysis. 314

315

Booster dose 316

A dose that is given at a certain time interval after completion of the primary series that is 317

intended to boost immunity to, and therefore prolong protection against, the disease that is to be 318

prevented. 319

320

Case ascertainment 321

The method adopted in a trial of vaccine efficacy for detecting cases of the infectious disease 322

intended to be prevented by vaccination. 323

324

Case definition 325

The pre-defined clinical and laboratory criteria that must be fulfilled to confirm a case of a 326

clinically manifest infectious disease in a study of vaccine efficacy or effectiveness. 327

328

Clinical trial application 329

An application submitted to a NRA by a sponsor for the purposes of gaining authorization to 330

conduct a clinical trial of an investigational or licensed vaccine at a trial site within the NRA’s 331

jurisdiction. The contents and format of the application will vary as required by the relevant 332

NRA(s). 333

334

Cluster randomization 335

Randomization of subjects into a clinical trial by group (e.g. by households or communities) as 336

opposed to randomization of the individual subject. 337

338

Geometric mean concentration 339

The average antibody concentration for a group of subjects calculated by multiplying all values 340


and taking the nth root of this number, where n is the number of subjects. 341

342

Geometric mean titre 343

The average antibody titre for a group of subjects calculated by multiplying all values and taking 344

the nth root of this number, where n is the number of subjects. 345

346

Good clinical practice (GCP) 347

GCP is a process that incorporates established ethical and scientific quality standards for the 348

design, conduct, recording and reporting of clinical research involving the participation of 349

human subjects. Compliance with GCP provides public assurance that the rights, safety, and 350

well-being of research subjects are protected and respected, consistent with the principles 351

enunciated in the Declaration of Helsinki and other internationally recognized ethical guidelines, 352

and ensures the integrity of clinical research data. 353

354

Good manufacturing practice (GMP) 355

GMP is the aspect of quality assurance that ensures that medicinal products are consistently 356

produced and controlled to the quality standards appropriate to their intended use and as required 357

by the product specification. 358

359

Immunological correlate of protection (ICP) 360

An Immunological Correlate of Protection (ICP) is most commonly defined as a type and 361

amount of immunological response that correlates with vaccine-induced protection against a 362

clinically apparent infectious disease and is considered predictive of clinical efficacy. For 363

some types of vaccines the ICP may be the type and amount of immunological response that 364

correlates with vaccine-induced protection against infection (e.g. hepatitis A and B vaccines). 365

The ICP may be mechanistic (i.e. causative for protection, such as antibody that effects virus 366

neutralization or serum bactericidal antibody) or it may be non-mechanistic (i.e. non-causative, 367

an immune response that is present in those protected by vaccination, but not the cause of 368

protection (such as serum IgG against VZV in the context of prevention of herpes zoster). 369

370

Immune memory 371


An immunological phenomenon in which the primary contact between the host immune system 372

and an antigen results in a T-cell-dependent immune response, often referred to as priming of the 373

immune system. Effective priming results in development of memory B-cells and an anamnestic 374

immune response to post-primary doses, which are commonly referred to as booster doses. 375

376

Immunogenicity 377

The capacity of a vaccine to elicit a measurable immune response. 378

379

Non-inferiority trial 380

In the context of vaccine clinical development programs, non-inferiority trials may have the 381

primary objective of showing that the immune response(s) to one or more specific antigenic 382

components in a candidate vaccine are not inferior to immune responses to corresponding 383

antigenic components in a licensed vaccine. Alternatively, the primary objective may be to 384

demonstrate that a candidate vaccine has non-inferior efficacy to a licensed vaccine. 385

386

Pharmacovigilance 387

A practice of detecting, assessing, understanding, responding to and preventing adverse drug 388

reactions, including reactions to vaccines, in the post-licensure period. 389

390

Posology 391

The vaccine posology for a specific route of administration and target population includes: 392

The dose content and volume delivered per dose 393

The dose regimen (i.e. the number of doses to be given in the primary series and, if 394

applicable, after the primary series) 395

Dose schedule (i.e. the dose intervals to be adhered to within the primary series and between 396

the primary series and any further doses) 397

398

Post-licensure safety surveillance 399

A system for monitoring AEFIs in the post-licensure period. 400

401

Post-primary doses 402


Additional doses of vaccine given after some time interval following the primary series of 403

vaccination, which may or may not boost the immune response. 404

405

Primary vaccination 406

First vaccination or series of vaccinations intended to establish clinical protection. 407

408

Protocol 409

A document that states the background, rationale and objectives of the clinical trial and describes 410

its designs, methodology and organization, including statistical considerations and the conditions 411

under which it is to be performed and managed. The protocol should be signed and dated by the 412

investigator, the institution involved and the sponsor. 413

414

Randomization 415

In its simplest form, randomization is a process by which n individuals are assigned to a test (nT) 416

or control (nC) treatment so that all possible groups of size n = nT + nC have equal probability of 417

occurring. Thus, randomization avoids systematic bias in the assignment of treatment. 418

419

Responder 420

A vaccinee who develops an immune response (humoral or cellular) that meets or exceeds a pre-421

defined threshold value using a specific assay. This term is most often used when there is no ICP 422

and when the clinical relevance of achieving or exceeding the pre-defined response is unknown. 423

424

Responder rate 425

The responder rate is the percentage of vaccinees achieving or exceeding the pre-defined level of 426

response. 427

428

Serious adverse event (SAE) or serious AEFI (SAEFI) 429

An adverse event is serious when it results in death, admission to hospital, prolongation of a 430

hospital stay, persistent or significant disability or incapacity, is otherwise life-threatening or 431

results in a congenital abnormality/birth defect. SAEs are such events that occur during clinical 432

trials. SAEFIs are such events that occur during post-licensure safety surveillance. 433


434

Seroconversion 435

A predefined increase in antibody concentration or titre. In subjects with no measurable antibody 436

prior to vaccination seroconversion is usually defined as achieving a measurable antibody level 437

post-vaccination. In subjects with measurable antibody prior to vaccination seroconversion is 438

commonly defined by a pre-defined fold-increase from pre- to post-vaccination. The definitions 439

may be adjusted depending on whether the lower limit of detection of the assay is or is not the 440

same as the lower limit of quantification. 441

442

Sponsor 443

The individual, company, institution or organization that takes responsibility for the initiation, 444

management and conduct of a clinical trial. The entity acting as a sponsor for a clinical trial is 445

usually the same as that which applies for clinical trial approval. The sponsor of a clinical trial 446

may not be the entity that applies for a license to place the same product on the market and/or the 447

entity that holds the license (i.e. is responsible for post-licensing safety reporting) in any one 448

jurisdiction. 449

450

Superiority trial 451

A trial with the primary objective of demonstrating that the immune response to one or more 452

antigenic components in a group that receives a candidate vaccine is superior to the 453

corresponding immune response in a control group. 454

455

Vaccine efficacy 456

An estimate of the reduction in the chance or odds of developing clinical disease after 457

vaccination relative to the chance or odds when not vaccinated against the disease to be 458

prevented. Vaccine efficacy measures direct protection (i.e. protection induced by vaccination in 459

the vaccinated population sample). 460

461

Vaccine effectiveness 462

An estimate of the protection conferred by vaccination in a specified population that measures 463

both direct and indirect protection (i.e. the estimate may reflect in part protection of non-464


vaccinated persons secondary to the effect of the vaccine in the vaccinated population). 465

466

Vaccine vector 467

A vaccine vector is a genetically engineered micro-organism (which may be replication 468

competent or incompetent) that expresses one or more foreign antigen(s) (i.e. antigens derived 469

from a different micro-organism). 470

471

472

4. Vaccine Clinical Development Programs 473

474

This Section considers: 475

Important considerations for clinical programs, including: 476

- Consultations with regulatory authorities 477

- Use of independent data review committees 478

- Registering and reporting clinical trials 479

Typical clinical development programs for new candidate vaccines, including: 480

- Main objectives of the clinical development program 481

- Factors that determine the extent and content of the program 482

- Stages of typical development programs 483

- Programs that do and do not include vaccine efficacy trials 484

- Alternatives for estimation of vaccine efficacy 485

Clinical evaluation trials after initial licensure 486

487


489

For a new candidate vaccine the main objective of the clinical development program is to 490

accumulate adequate data to support initial licensure and appropriate use, as described in 491

Subsection 4.2. The essential elements of the program are: 492

To describe the interaction between the vaccine and the host immune response (Section 5) 493

To identify safe and effective dose regimens and schedules (Sections 5 and 6) 494

To provide estimates of vaccine efficacy by directly measuring efficacy or inferring efficacy 495


based on immune responses (Sections 5 and 6) 496

To describe the safety profile (Section 7) 497

To assess co-administration with other vaccines if this will be essential for use (Section 5) 498

499

After initial licensure, as described in Subsection 4.3: 500

It is essential to monitor vaccine safety in routine use (Section 7). 501

It is commonly appropriate to estimate vaccine effectiveness (Section 6) 502

Depending on the content of the pre-licensure program, further trials of safety, 503

immunogenicity and/or efficacy may be conducted and the data may be used to extend or 504

otherwise modify the use of the vaccine via amendment of the prescribing information. 505

506

4.1.1 Consultation with National Regulatory Authorities (NRAs) 507

508

It is strongly recommended that dialogue with the appropriate NRAs occurs at regular intervals 509

during the pre-licensure clinical development program to agree on the content and extent of the 510

initial application dossier. This is especially important when: 511

a. The clinical program proposes a novel approach to any aspect of development for which 512

there is no precedent or guidance available 513

b. The proposed program conflicts with existing guidance to which the NRAs involved would 514

usually refer when considering the suitability of the program 515

c. There are particular difficulties foreseen in providing evidence to support an expectation of 516

vaccine efficacy (i.e. there is no immunological correlate of protection and a vaccine 517

efficacy study is not feasible) 518

d. There are other special considerations for the total content of the pre-licensure program. For 519

example, when it is necessary to use different vaccine constructs for priming and boosting to 520

achieve immune responses thought likely to be protective. In this case each constitutes a 521

separate vaccine but the clinical data required to support their licensure for use in tandem is 522

less than would be required for two vaccines intended to be used completely independently. 523

524

Further dialogue should ensue whenever additional clinical trials are planned with intent to 525

modify the prescribing information. In addition, it should be considered whether changes to the 526


manufacturing process of a vaccine before or after initial licensure need to be discussed with 527

NRAs to establish whether or not specific clinical trials are required to support the changes. 528

Consultation with NRAs is also essential when issues of vaccine safety or effectiveness arise in 529

the post-licensure period to determine any actions that are needed. 530

531

4.1.2 Use of independent data monitoring committees 532

533

It is common in vaccine trials that a data safety monitoring board (DSMB) is appointed to 534

provide independent ongoing assessments of safety data. In the pre-licensure program for a new 535

candidate vaccine it may be appropriate to have a DSMB in place even for the initial exploratory 536

trials and dose-finding trials, especially if the vaccine consists of a new construct and/or when it 537

may be anticipated that it could be very reactogenic. For other vaccines it may be considered 538

useful to have a DSMB in place if available data from the same or similar vaccines point to the 539

possibility of important safety issues or if the trial will enrol particular populations (e.g. infants 540

and toddlers, pregnant women or immunocompromised subjects). A DSMB may not be 541

considered necessary for trials with vaccines that include only established antigenic components 542

and adjuvants for which no particular safety problems are anticipated or when a licensed vaccine 543

is being investigated using an alternative posology or in a new population. If the DSMB charter 544

includes recommending that trials are terminated early for safety reasons there should be 545

appropriate stopping rules in place. 546

547

In vaccine efficacy trials it may also be appropriate to appoint an independent data adjudication 548

committee consisting of individuals with expertise relevant to the infectious disease to be 549

prevented. For example, such a group could be used to provide an independent review of the 550

eligibility of individual vaccinees for inclusion in the primary analysis population and/or to 551

identify cases of clinically apparent infections that meet the pre-defined case definition. If such a 552

committee is appointed to oversee one or more trials the protocol and statistical analysis plan 553

should clarify whether the conclusions of the adjudication committee will be used to conduct the 554

primary analysis and any secondary analyses that are pre-defined. 555

556

In some situations, it may be appropriate to appoint an independent data monitoring committee 557


to review the results of pre-planned interim analyses of safety and/or efficacy data when a certain 558

proportion of the intended sample size has reached a certain stage of participation. It may be 559

appropriate that the DSMB or some other independent data monitoring committee takes on this 560

responsibility. Protocols and statistical analysis plans may define futility criteria to be applied to 561

the results of one or more interim analyses that, if met, would result in a recommendation from 562

the independent committee to terminate the trial. Whenever an interim analysis is planned, expert 563

statistical input should be obtained to ensure that appropriate adjustments are made to protect the 564

power and integrity of the trial. 565

566

4.1.3 Registering and reporting clinical trials 567

568

Before any clinical trial is initiated (i.e. before the first subject receives the first medical 569

intervention in the trial) its details must be registered in a publicly available, free to access, 570

searchable clinical trial registry. The registry should comply with individual NRA requirements 571

and as a minimum should comply with the WHO international agreed standards. 572

573

The entry into the clinical trial registry site should be updated as necessary to include final 574

enrolment numbers achieved and the date of actual study completion (i.e. the last data collection 575

time point for the last subject for the primary outcome measure). If clinical trials are terminated 576

prematurely the entry should be updated to reflect this with a report of the numbers enrolled up 577

to the point of termination. 578

579

The key outcomes of a clinical trial must be posted in the results section of the entry in the 580

clinical trial registry within 12 months of study completion and/or posted on a publicly-available, 581

free-to-access, searchable website (e.g. that of the trial sponsor or Principal Investigator). 582

583

Each NRA may have specific requirements for reporting the results of completed trials and the 584

status of ongoing clinical trials conducted with a specific product within and without their 585

jurisdiction. Whatever these requirements, each regulatory submission (whether for clinical trial 586

approval, to support initial licensure or a post-licensure modification or to provide a product 587

safety update report) should include a listing of all completed and ongoing trials conducted with 588


the product by the sponsor. It is recommended that any trials that are known to the sponsor (e.g. 589

from searching registries or from publications) that were initiated by persons other than the 590

sponsor (e.g. by a public health body or academic institution or by another company that used 591

the product as a comparator) should also be listed. 592

593

4.2 New candidate vaccines 594

595

Examples of new candidate vaccines from the regulatory standpoint include: 596

i. Vaccines that contain only new antigenic components (i.e. not previously used in 597

licensed vaccines) 598

ii. Vaccines that contain both new (i.e. not in any licensed vaccine) and known (i.e. already 599

in licensed vaccines) antigenic components 600

iii. Vaccines that contain a new adjuvant, with known and/or new antigenic components 601

iv. Vaccines that contain only known antigenic components that have not previously been 602

combined all together into a single vaccine, with or without a known adjuvant 603

v. Vaccines that contain only known antigenic components ± known adjuvants in a 604

combination that is already licensed but the vaccine is produced by a different 605

manufacturer. This includes situations in which seed lots or bulk antigenic components 606

used to make a licensed vaccine are supplied to other manufacturers for their own vaccine 607

production. 608

609

For new candidate vaccines the content and extent of pre-licensure clinical development 610

programs will reflect how much is already known about the antigenic components and adjuvants 611

in the product. Some of the most important factors include: 612

a. Number of the antigenic components (e.g. from the same or from several infectious 613

organisms) 614

b. Nature of the antigenic components (e.g. manufactured with or without genetic 615

modification, live attenuated, live vectored) 616

c. Inclusion of an adjuvant 617

d. Disease(s) to be prevented 618

e. The available options for predicting vaccine efficacy (e.g. inferring efficacy based on 619


established immunological correlates of protection or conducting vaccine efficacy trials) 620

f. Age range and population for use (e.g. infants, elderly, pregnant women) 621

g. Route of administration 622

h. Likelihood of co-administration with other vaccines in routine use 623

i. Vaccine-specific safety issues that may be anticipated 624

625

4.2.1 Safety and immunogenicity trials 626

627

The safety and immunogenicity of a new candidate vaccine should be evaluated in all pre-628

licensure clinical trials. In the earliest stage of clinical development the primary objective of a 629

trial is usually to describe safety although immunogenicity data are also collected. In later trials 630

the primary objective is usually to address specific immunogenicity issues and the assessment of 631

safety may be a co-primary or secondary objective. In vaccine efficacy trials evaluations of 632

safety and immunogenicity are usually secondary objectives (see Subsection 4.2). 633

634

4.2.1.1 Initial trials 635

636

These are commonly referred to as Phase 1 trials. 637

638

The clinical program for new candidate vaccines commences with an exploration of safety and 639

of the interaction between the antigens proposed for inclusion in the candidate vaccine and the 640

human immune system. In most cases the first clinical trials are conducted in healthy young 641

adults before proceeding to conduct trials in other age groups and/or in subjects with underlying 642

conditions. Depending on the perceived benefit and risks of vaccination it may not be 643

appropriate or necessary to apply an age de-escalation approach (e.g. to move from adults to 644

adolescents, then to children aged 6-12 followed by younger children, toddlers and finally 645

infants) to sequential trials or groups within trials. For example, if a vaccine has negligible 646

potential benefit for older children it may be acceptable in some cases to proceed from trials in 647

adults to trials in infants and toddlers. 648

649

It is usual that these trials explore different doses of antigenic components and, if applicable, the 650


effect of adding an adjuvant in various amounts. For vaccines that contain more than one new 651

antigenic component the first trials may evaluate each one given alone before selecting possible 652

doses for use in combinations. When new antigenic components are to be added to a licensed 653

product the immune response to separate administrations and to the proposed combination 654

product are compared. For vaccines that contain only known antigenic components and 655

adjuvants the initial trials focus on the effects of combining them into a single formulation or the 656

effects of mixing immediately prior to injection (e.g. using a liquid formulation of some 657

component to reconstitute a lyophilized presentation of the others). Depending on the initial 658

results, sequential trials may explore formulations with adjusted amounts of one or more 659

antigenic components and/or the adjuvant. 660

661

4.2.1.2 Further trials 662

663

These are commonly referred to as Phase 2 trials. 664

665

Further safety and immunogenicity trials are conducted to build on the Phase 1 trial results. In 666

most cases these trials are conducted in subjects who are representative of the intended target 667

population for the vaccine at the time of initial licensure. 668

669

These trials are usually designed to provide sufficient immunogenicity data to support selection 670

of one or more candidate formulations for further trial i.e. to select the amounts of antigenic 671

components and, where applicable, adjuvants in each dose. They may provide adequate data to 672

determine the number of doses and dose intervals but the final vaccine posology is sometimes 673

established only after completion of confirmatory immunogenicity trials or vaccine efficacy 674

trials. 675

676

4.2.1.3 Confirmatory (or pivotal) trials 677

678

In many vaccine clinical development programs the confirmatory (or pivotal) trial(s) involve an 679

estimate of vaccine efficacy as described in Subsection 4.2.2. 680

681


In instances where vaccine efficacy trials do not need to be, or cannot be, conducted (see 682

Subsection 4.2.2), the confirmatory (or pivotal) trial(s) usually assess the immunogenicity of the 683

final selected vaccine formulation and posology in each target population. In this setting, they 684

are commonly referred to as Phase 3 safety and immunogenicity trials. It is usual that the 685

investigational formulations used in these confirmatory safety and immunogenicity trials (as well 686

as in confirmatory efficacy trials; see below) should be manufactured using validated processes 687

and should undergo lot release in the same way as intended for the commercial product. 688

689

4.2.2 Efficacy trials 690

691

Vaccine efficacy trials have the primary aim of evaluating the protective efficacy of a candidate 692

vaccine against an infectious disease. The immunogenicity data collected during vaccine efficacy 693

trials can be used to evaluate the relationship between immune parameters and efficacy and may 694

enable identification of immune correlates of protection (see Subsection 5.4). These trials also 695

provide an opportunity to collect extensive safety data using the final intended formulation and 696

dose regimen in the target population. 697

698

Preliminary vaccine efficacy trials may be conducted to explore the magnitude of protection that 699

may be possible and to inform the design of confirmatory vaccine efficacy trials (e.g. by 700

evaluating efficacy of different dose regimens and/or by estimating efficacy based on a range of 701

efficacy variables). If conducted, these are commonly referred to as Phase 2b trials. They are also 702

sometimes referred to as pilot efficacy trials or proof of concept efficacy trials. 703

704

Confirmatory vaccine efficacy trials that are designed and powered to provide statistically robust 705

estimates of vaccine efficacy are commonly referred to as Phase 3 (or pivotal) efficacy trials or 706

sometimes as field efficacy trials. 707

708

The need for and feasibility of evaluating the protective efficacy of a candidate vaccine should 709

be considered at an early stage of vaccine development because the conclusion will determine 710

the overall content of the pre-licensure clinical program and impact on its duration. In all 711

application dossiers that do not include an evaluation of vaccine efficacy the sponsor should 712


provide a sound justification for the lack of such data, taking into account the following: 713

714

a) Efficacy data are not required 715

716

Vaccine efficacy trials are not necessary if it is established that clinical immunological data can 717

be used to predict protection against disease. For example, when there is an established 718

immunological correlate for protection against a specific disease (e.g. anti-toxin levels against 719

diphtheria and tetanus toxins, antibody against hepatitis B surface antigen) the candidate vaccine 720

should be shown to elicit satisfactory responses based on the relevant correlate(s). 721

722

b) Efficacy data are usually required 723

724

Vaccine efficacy trials are usually required whenever a candidate vaccine is developed with 725

intent to protect against an infectious disease and one or more of the following apply: 726

There is no established immunological correlate of protection that could be used to predict 727

the efficacy of the candidate vaccine. 728

There is no existing licensed vaccine of documented efficacy against a specific infectious 729

disease to allow for immunobridging of a candidate vaccine to the efficacy of a licensed 730

vaccine. 731

Immunobridging to the documented efficacy of a licensed vaccine against a specific 732

infectious disease is not considered to be possible because there is no known relationship 733

between specific immune response parameters and efficacy. 734

There are sound scientific reasons to expect that vaccine efficacy cannot be extrapolated 735

from the population(s) included in the prior efficacy trial(s) with a candidate vaccine to one 736

or more other populations. 737

There are sound scientific reasons to expect that vaccine efficacy that has been demonstrated 738

for the candidate vaccine against infectious disease due to specific strains (e.g. serotypes, 739

sub-types) cannot be extrapolated to other strains. 740

741

c) Efficacy data cannot be provided 742

743


In some instances in which efficacy data are usually required it may not be feasible to conduct 744

efficacy trials. For example, if the candidate vaccine is intended to prevent an infectious disease 745

that: 746

o Does not currently occur (e.g. smallpox) 747

o Occurs in unpredictable and short-lived outbreaks that do not allow enough time for the 748

conduct of appropriately designed trials to provide a robust estimation of vaccine efficacy 749

(e.g. some viral haemorrhagic fevers) 750

o Occurs at a rate that is too low for vaccine efficacy to be evaluated in a reasonably sized trial 751

population and period of time. This situation may apply: 752

a. Due to natural rarity (e.g. plague, anthrax, meningitis due to N. meningitidis type B) of 753

the infectious disease 754

b. Due to rarity of the infectious disease resulting from the widespread use of effective 755

vaccines. In this case the numbers required to conduct an adequately powered analysis 756

of the relative efficacy of a candidate vaccine vs. a licensed vaccine may be too large to 757

permit completion in any reasonable timeframe. 758

c. When the aim is to evaluate vaccine efficacy against serotypes or subtypes of an 759

organism that occur rarely (e.g. pneumococcal conjugate vaccines and human 760

papillomavirus vaccines). 761

762

If it is not feasible to perform vaccine efficacy trials and there is no immunological correlate of 763

protection, it may be possible to support an assumption of the likely efficacy of a vaccine by 764

deriving a marker of protection from one or more of the following: 765

i) Nonclinical efficacy trials 766

ii) Passive protection trials (i.e. effects of normal or hyper-immune human gamma 767

globulin, use of convalescent sera) that may point to the sufficiency of humoral 768

immunity for prevention of clinical disease and suggest a minimum protective antibody 769

level that could be used as a benchmark in clinical trials with candidate vaccines 770

iii) Trials of the acquisition of natural immunity that may support an approach as in ii) 771

iv) Human challenge trials 772

v) Comparison of immunological responses with those seen in past trials of similar 773

vaccines with proven protective efficacy (e.g. acellular pertussis vaccines) even though 774


the relationship between immune responses to one or more antigenic components and 775

efficacy remains unknown 776

777

4.2.3 Pivotal safety trials 778

779

Safety is an important secondary endpoint in all trials with the primary objective of assessing 780

immunogenicity or efficacy. In rare cases, the assessment of safety may be the primary or co-781

primary objective in a pre-licensure Phase 3 (pivotal trial) that has immunogenicity and/or 782

efficacy as secondary objectives, as described in Subsection 7.2.3. 783

784

4.3 Post-licensure clinical evaluations 785

786

For all licensed vaccines safety data are collected as part of routine pharmacovigilance. On 787

occasion, additional pharmacovigilance in the form of trials designed to address specific safety 788

issues that were identified as potential concerns from pre-licensure trials may be conducted 789

post-licensure (see Section 7). 790

791

Whether or not vaccine efficacy trials were conducted prior to initial licensure it is usual to 792

evaluate vaccine effectiveness during routine use or by means of trials specifically designed to 793

provide estimates of effectiveness (see Subsection 6.3). 794

795

Further clinical trials are commonly conducted after first licensure and are sometimes performed 796

to address commitments made to NRAs. These trials may or may not be intended to support 797

modifications of the prescribing information and may include: 798

a. Extension phases of trials that commenced before first licensure (e.g. to continue follow-up 799

of safety, efficacy and/or immune response, to evaluate the effects of further doses) 800

b. Trials that evaluate the use of alternative dose regimens (e.g. reducing the number of doses) 801

and/or schedules (e.g. extending the interval between doses) 802

c. Trials in additional populations (e.g. different age groups, populations with factors that 803

could affect their immune response, such as pregnancy, prematurity and 804

immunosuppression) 805


d. Trials to support changes in vaccine manufacture with potential to affect safety, efficacy or 806

immune response 807

e. Trials to support co-administration with other vaccines 808

809

The nomenclature for these types of trial is variable. If these additional trials are conducted in 810

wholly new populations or with substantially different vaccination regimens, especially when 811

they are intended to provide support for changes to the prescribing information, they are 812

commonly referred to as Phase 2 or 3 trials. Trials that are intended to support more minor 813

changes, such as adding alternative dose regimens or extending the age range, are commonly 814

referred to as Phase 3b trials. Other types of post-licensure trials, such as those in which vaccines 815

are given in accordance with licensed uses and regimens, are more often referred to as Phase 4 816

trials. These include trials that are specifically designed to address specific safety issues or to 817

estimate vaccine effectiveness. 818

819

5. Immunogenicity 820

821


The range of immunogenicity data that may be collected throughout the pre- and post-823

licensure clinical development program 824

Collection of specimens for immunogenicity trials 825

Characterization of the immune response to a new candidate vaccine 826

Selection of the immune parameters to be measured 827

Assays for measuring humoral and cellular immune responses 828

Identification and uses of immunological correlates of protection 829

Objectives and designs of immunogenicity trials 830

Considerations for some specific types of immunogenicity trials, including: 831

- Trials to identify formulations and posologies (primary and post-primary) 832

- Comparative immunogenicity trials to bridge efficacy 833

- Trials to extend or modify use 834

- Co-administration trials 835

- Trials in which pregnant women are vaccinated 836


- Trials to support major changes to the manufacturing process 837

- Lot to lot consistency trials 838

839


841

Immunogenicity trials are conducted at all stages of pre-licensure vaccine development and 842

additional trials are commonly conducted in the post-licensure period. In all trials the evaluation 843

of immune responses rests on the collection of adequate specimens at appropriate time intervals 844

and measurement of immune parameters most relevant to the vaccine using validated assays. 845

846

In the clinical development program for new candidate vaccines that contain micro-organisms 847

or antigens not previously included in human vaccines immunogenicity trials should provide a 848

detailed understanding of the immune response to vaccination. Subsequent pre-licensure and 849

post-licensure clinical trials commonly evaluate and compare immune responses between trial 850

groups to address a range of objectives. Depending on the objectives, stage of development and 851

trial population the comparisons may be made with one or more of placebo, other formulations 852

or regimens of the same vaccine or licensed vaccines. In these trials the assessments and 853

analyses of the immune responses are primary objectives whereas the assessments of safety 854

may be co-primary or secondary objectives. In trials that are primarily intended to estimate 855

vaccine efficacy, assessment of the immune responses is usually a secondary objective but it is 856

important that data on immune responses are collected to support analyses of the relationship 857

between immunogenicity and efficacy, which may lead to identification of immunological 858

correlates of protection. 859

860

5.2 Characterization of the immune response 861

862

For micro-organisms and antigens that have not been used previously in human vaccines a 863

thorough investigation of their interaction with the human immune response should be conducted 864

as part of the overall clinical development program. For micro-organisms and antigens that are 865

already in licensed vaccines it is not usually necessary to repeat these types of investigations but 866

consideration should be given to conducting at least some trials in certain circumstances (e.g. 867


when a new adjuvant is to be added to known antigens, a different method of attenuation is used, 868

a different carrier protein is used for antigen conjugation or an antigen previously obtained by 869

purification from cultures is to be manufactured using recombinant technology). 870

871

The range of investigations conducted should take into account what is known about the immune 872

response that results from natural exposure and whether or not this provides partial or complete 873

protection that is temporary or lifelong. The range of investigations should also consider the 874

characteristics of the infecting micro-organism (e.g. whether there are multiple subtypes that 875

cause human disease) and the content of the vaccine (14). Investigations may include some or all 876

of the following: 877

Determination of the amount, class, sub-class and function of antibody elicited by the 878

vaccine 879

Description of the magnitude of the humoral and cell-mediated immune response to initial 880

and sequential doses and changes in the magnitude of responses with time elapsed since 881

vaccination 882

Assessment of the ability of the vaccine to elicit a T-cell dependent primary immune 883

response, with induction of immune memory (i.e. priming of the immune system) giving rise 884

to anamnestic responses i) on natural exposure ii) after further doses of the same vaccine 885

and/or iii) after further doses of a vaccine that contains closely related but non-identical 886

micro-organisms or antigens (i.e. cross-priming) 887

Assessment of the specificity and cross-reactivity of the immune response 888

Assessment of changes in antibody avidity with sequential doses, which may be useful when 889

investigating priming 890

Evaluation of factors that could influence the immune responses (e.g. presence of maternal 891

antibody, pre-existing immunity to the same or very similar organisms, natural or vaccine-892

elicited antibody against a live viral vector) 893

894

5.3 Measuring the immune response 895

896

5.3.1 Collection of specimens 897

898


Immune responses to vaccination are routinely measured in serum (humoral immune responses) 899

and blood (cellular immune responses). For some vaccines it may be of interest to explore 900

immune responses in other body fluids that are relevant to the site at which the target micro-901

organism infects and/or replicates (e.g. in nasal washes or cervical mucus), especially if it is 902

known or suspected that the systemic immune response does not show a strong correlation with 903

protective efficacy for the type of vaccine under trial (e.g. intranasal vaccination against 904

influenza). Nevertheless, to date specimens other than sera have not provided data that have been 905

pivotal in regulatory decision making processes and have not resulted in identification of ICPs. 906

Therefore the rest of this section focuses on the collection of sera. 907

908

Pre-vaccination samples should be collected from all subjects in the early immunogenicity trials 909

after which it may be justifiable to omit these samples or to obtain them from subsets (e.g. if the 910

initial trials indicate that antibody is rarely detectable or quantifiable prior to vaccination in the 911

target population). Pre-vaccination sampling remains essential if it is expected that the target 912

population will have some degree of pre-existing immunity either due to natural exposure and/or 913

their vaccination history since the assessment of the immune response will need to take into 914

account seroconversion rates and increments in geometric mean titres or concentrations from 915

pre- to post-vaccination. Pre-vaccination sampling is also necessary if it is known or suspected 916

that pre-existing immune status may have a positive (e.g. because pre-existing antibody reflects 917

past priming) or negative (e.g. due to maternal antibody interfering with primary vaccination 918

with certain antigens in infants) impact on the magnitude of the immune response to vaccination. 919

920

The timing of post-vaccination sampling should be based on what is already known about the 921

peak immune response and antibody decay curve after initial and, if applicable, sequential doses 922

(e.g. for vaccines that elicit priming the rise in antibody after a booster dose is usually much 923

more rapid compared to earlier doses). For antigens not previously used in human vaccines 924

sampling times may be based initially on nonclinical data and then adjusted when antibody 925

kinetic data specific to the antigen(s) under trial have been generated. As information is 926

accumulated the number and volume of samples taken from individual vaccinees may be reduced 927

to the minimum considered necessary to address the trial objectives. 928

929


5.3.2 Immunological parameters 930

931

Immunological parameters are measures that describe the humoral (e.g. antibody concentrations 932

or antibody titres depending on the assay output) or the cell-mediated (e.g. percentages of 933

sensitised T-cells) immune response. To date, immunological parameters other than those that 934

measure the humoral immune response have not played a pivotal or major role in vaccine 935

licensure so that the focus is usually on determination of antibody levels. 936

For known micro-organisms or antigens in a candidate vaccine the range of parameters to be 937

measured in clinical trials is usually selected from prior experience and whether or not there 938

is an established ICP. 939

For micro-organisms or antigens not previously included in human vaccines the selection of 940

parameters to be measured should take into account what is known about natural immunity. 941

For some infectious diseases the nature of the immune response to infection in animal 942

models may also be useful for parameter selection. In later clinical trials, after 943

characterization of the immune response, the parameters to be measured may be modified. 944

945

5.3.2.1 Humoral immune response 946

947

The humoral immune response is assessed from the post-vaccination appearance or increase 948

from pre-vaccination in antibody directed at specific micro-organisms or antigens in the vaccine. 949

Most weight is usually placed on functional antibody responses (e.g. serum bactericidal 950

antibody [SBA], toxin or virus neutralizing antibody, opsonophagocytic antibody [OPA]) 951

but there may not be an appropriate assay available (e.g. for typhoid vaccines based on the 952

Vi polysaccharide) or the only available assays may have low feasibility for application to 953

large numbers of samples (e.g. because they are very labor intensive or require high-level 954

biocontainment facilities). 955

Alternatively, or in addition to the determination of functional antibody, the immune 956

response may be assessed by measuring total antibody (e.g. total IgG measured by ELISA) 957

that binds to selected antigens (or, on occasion, to specific epitopes). Only a proportion of 958

the total antibody detected may be functional. 959

960


The following should be taken into consideration when deciding how to measure the humoral 961

immune response: 962

a. If a strong correlation has already been established between total and functional antibody 963

responses to a specific micro-organism or antigen it may be acceptable to measure only total 964

IgG in further trials (e.g. antibody to tetanus toxin) 965

b. For antigens for which there is an established ICP it may suffice to measure only the 966

relevant functional antibody (e.g. SBA for meningococcal vaccines) or total IgG (e.g. for 967

antibody to tetanus toxin) response 968

c. If the ICP is based on total IgG there may be instances in which there is still merit in 969

measuring functional antibody (e.g. for antibody to diphtheria toxin for which a micro-970

neutralization assay is available) 971

d. If there is no ICP the functional antibody response should be measured if this is feasible 972

e. Occasionally there may be more than one immunological parameter that measures functional 973

antibody but one is considered to be a more definitive measure than the other (e.g. 974

neutralizing antibody to influenza virus vs. antibody that inhibits haemagglutination), in 975

which case the more definitive parameter may be determined at least in a subset 976

f. For some vaccines against certain viruses there is a potential that some of the total antibody 977

detected has no protective effect (e.g. is non-neutralizing) but it could enhance cellular 978

infection by wild-type virus and result in an increased risk of severe disease after 979

vaccination (e.g. this may apply to dengue vaccines). To assess this possibility the routine 980

measurement of total antibody to assess the humoral immune response to vaccination should 981

be supported by other detailed investigations. 982

983

5.3.2.2 Cell-mediated immune response 984

985

For some types of infectious disease (such as tuberculosis) the assessment of the cell-mediated 986

immune response may have a major role in the assessment of the interaction between the 987

vaccine and the human immune system. In many other settings the evaluation of the cellular 988

immune response may serve to support the findings based on the humoral immune response 989

(e.g. when assessing the benefit of adding an adjuvant or when evaluating the degree of cross-990

priming elicited by a vaccine). 991


992

The cell-mediated immune response is most commonly assessed by detecting and quantifying 993

sensitized T-cells in blood from vaccinees. These investigations may also serve to characterize 994

the predominant cytokines released and to detect differences in sensitization between T-cell sub-995

populations. There are several methods that may be used. These are commonly based on 996

measuring the production of a range of cytokines following in-vitro stimulation of T-cells with 997

individual or pooled antigens. 998

999

To date, the methodologies used for these and alternative types of assays have been variable and 1000

non-standardized. Nevertheless, the results may provide useful comparisons between treatment 1001

groups within any one study (e.g. could describe the effect, if any, of an adjuvant) based on 1002

comparing rates of “responders” defined by a magnitude of change in the assay readout from 1003

pre- to post-vaccination. If there are marked discrepancies in the patterns of responses observed 1004

between cell-mediated and humoral responses (e.g. if adding an adjuvant does have a major 1005

effect on antibody levels but does not increase the percentages of sensitized cells in one or more 1006

T-cell subsets) the findings should be carefully considered and discussed. 1007

1008

5.3.3 Assays 1009

1010

Assays of functional or total antibody that are used to report immune responses to vaccination 1011

(whether to the candidate vaccine or to co-administered vaccines) in trials intended to support 1012

licensure (i.e. in confirmatory trials) may be: 1013

Commercially available assays specifically designed and intended for quantification of 1014

antibody that are considered acceptable to NRAs (i.e. have been marketed following a robust 1015

regulatory review by the same or by other NRAs). 1016

In-house assays that have been validated according to similar principles recommended for 1017

quantitative lot release assays in the ICH Q2 (R1) document Validation of Analytical 1018

Procedures: Text and Methodology (15). In-house assays that are used in early trials that 1019

explore the immune response may be regarded as an exception and may report data using 1020

assays that have yet to be validated or which are not subsequently validated. 1021

In-house assays that have been shown to be comparable to a reference assay (e.g. to an assay 1022


established in a WHO reference laboratory or to an assay that is established in a recognized 1023

public health laboratory and which has been used previously to support clinical trials that 1024

have been pivotal for licensure). 1025

In each case, it is expected that WHO International Standard reagents will be used in assay runs 1026

if these exist or omission of their use should be adequately justified. 1027

1028

Commercial assays suitable for quantification of the cell-mediated response to vaccination are 1029

not currently available but may be used in future. In-house assays that are used to detect and 1030

quantify cell-mediated immunity may be difficult to fully validate, in which case the results 1031

should not be used to make specific claims regarding clinical effect. 1032

1033

Clinical trial protocols should specify which assays will be used and in which laboratories. 1034

Clinical trial reports should include at least a summary of the assay methodology and its 1035

commercial or other validation status. For in-house assays the validation reports should be 1036

provided. 1037

1038

It is preferable that the same assays are used in the same laboratories throughout the clinical 1039

development program (including pre-and post-licensure trials) for an individual vaccine. It is 1040

also preferable that each assay (whether it measures the response to the candidate vaccine or to a 1041

concomitant vaccine) is run by one central laboratory. If this is not possible (e.g. because 1042

different laboratories have to be used, commercial or in-house assays change over time or a 1043

switch is made between in-house and commercial assays) the new and original assays should be 1044

shown to be comparable. As a minimum it is recommended that a selection of stored sera (e.g. 1045

covering a range of low to high results when using the previous assay) are re-run using the 1046

previous and new assays in parallel. The number of sera re-tested should be sufficient to support 1047

a statistical assessment of inter-assay variability. 1048

1049

The micro-organisms (e.g. in assays of SBA, OPA and virus neutralization) and the antigens 1050

(e.g. in ELISAs and for in-vitro stimulation of sensitized T-cells) used in the assay may affect 1051

both the result and the interpretation of the result. For example: 1052

It is important to use purified antigen to avoid the possibility that the assay detects and 1053


measures antibody to any extraneous antigenic substances that may be in the vaccine. 1054

For vaccines that contain antigens from multiple strains of the same species (e.g. multiple 1055

bacterial capsular types) separate assays are needed to determine the immune response to 1056

each antigen. 1057

Although it is usually acceptable to conduct routine testing using the same micro-organisms 1058

or antigens present in the vaccine it may be very informative to perform additional testing, at 1059

least in subsets of samples, using circulating wild-type organisms or antigens derived from 1060

them in the assay. It is not expected that these additional assays will necessarily be validated 1061

since they are exploratory in nature. The results of additional testing can provide an 1062

indication as to whether the results of routine testing could represent an over-estimate of the 1063

immune response to circulating strains. This additional testing can also provide an 1064

assessment of the cross-reactivity of the immune responses elicited by the vaccine to other 1065

organisms of the same genus or species (e.g. to different flaviviruses, to different clades of 1066

influenza virus or to different HPV types) and guide the need to replace or add strains or 1067

antigens in a vaccine to improve or maintain its protective effect. 1068

1069

5.4 Identification and use of immunological correlates of protection 1070

1071

5.4.1 Immunological correlates of protection and their uses 1072

1073

To date, all established ICPs are based on humoral immune response parameters that measure 1074

functional or total IgG antibody. Examples of well-established ICPs include those for antibody 1075

to diphtheria and tetanus toxoids, polioviruses, hepatitis B virus and H. influenzae type b (Hib) 1076

polysaccharide (PRP) (16). In most cases, established ICPs have been shown to correlate with 1077

prevention of clinically apparent infectious disease but for some pathogens the ICP correlates 1078

with prevention of documented infection (e.g. hepatitis A and hepatitis B). 1079

1080

In some cases the ICP is a measure of the functional antibody response but if a strong correlation 1081

is shown between the results of assays of functional and total antibody, it may be possible to 1082

derive an alternative ICP based on total antibody (see Subsection 5.3.3). 1083

1084


Subsections 5.5.2 and 5.5.3 consider trial endpoints and the approach to analysis and 1085

interpretation of immunogenicity data in the presence or absence of an ICP and situations in 1086

which alternative approaches may be appropriate. For example, for some infectious diseases 1087

vaccine-elicited protection against clinical disease shows a broad correlation with a specific 1088

immunological parameter (e.g. with serum neutralising antibody elicited by HPV vaccines) but 1089

no cut-off value has been identified that shows a strong statistical correlation with protection in 1090

the short or longer-term in individuals or populations. In some other instances there is an 1091

indication of a threshold value that seems to broadly predict protection but the evidence is 1092

insufficient to regard this as an ICP applicable to a specific or to several different sub-1093

populations or organism subtypes (e.g. IgG to specific pneumococcal serotypes). For some other 1094

infectious diseases there is no correlation that is well established between vaccine-elicited 1095

protection and measurable immune parameters (e.g. for acellular pertussis vaccines). 1096

1097

5.4.2 Establishing an ICP 1098

1099

Documentation of the immune response to natural infection, the duration of protection after 1100

clinically apparent infection (i.e. whether natural protection is life-long [solid immunity], 1101

temporary or absent) and the specificity of protection (i.e. whether the individual is protected 1102

only against specific subtypes of a micro-organism) should be taken into account when 1103

attempting to establish an ICP from clinical data. For example, to date, widely-accepted clinical 1104

ICPs have been established based on one or more of: 1105

Serosurveillance and disease prevalence in specific populations 1106

Passive protection using antibody derived from immune humans or manufactured using 1107

recombinant technology 1108

Efficacy trials 1109

Effectiveness trials 1110

Investigation of vaccine failure in immunosuppressed populations 1111

1112

In the majority of cases clinical ICPs have been determined from vaccine efficacy trials that were 1113

initiated pre-licensure, often with long-term follow-up of subjects that extended into the post-1114

licensure period. Efficacy trial protocols should plan to collect sufficient information to allow for 1115


analyses of the relationship between immune parameters and protection against clinically 1116

apparent disease. As a minimum this requires collection of post-vaccination samples from all or 1117

from a substantial subset of the vaccinated and control groups. Serial collection of samples over 1118

the longer-term along with follow-up surveillance for vaccine breakthrough cases has also served 1119

to support identification of ICPs. 1120

1121

To investigate the predictive capacity of a putative ICP protocols should pre-define the 1122

assessments to be applied to all cases of the disease to be prevented that occur in the vaccinated 1123

and control groups. These assessments should include investigation of the immune status of 1124

subjects and microbiological studies with the infecting micro-organisms whenever these have 1125

been recovered. For breakthrough cases from which there are both post-vaccination sera and 1126

organisms recovered it is recommended that functional antibody should be determined (or, if not 1127

possible, total antibody) for individuals against their own pathogen. An exploration of vaccine-1128

elicited cell-mediated responses in individuals against their own pathogen may also be useful 1129

and, for some types of infectious diseases (such as tuberculosis), may be very important to 1130

further understanding of vaccine-associated protection. These data may be very important to 1131

investigate the broad applicability of the ICP depending on host and organism factors. 1132

1133

A single clinical ICP identified from a vaccine efficacy trial in a defined population may not 1134

necessarily be applicable to other vaccine constructs intended to prevent the same infectious 1135

disease. In addition, an ICP may not be applicable to other populations and disease setting. For 1136

example, putative ICPs have sometimes differed between populations of different ethnicities 1137

with variable natural exposure histories for subtypes of a single micro-organism. Thus the 1138

reliance that is placed on a clinical ICP, even if regarded as well-supported by the evidence, 1139

should take into account details of the efficacy trials from which it was derived. 1140

1141

Clinical ICPs have also been derived from or further supported by analyses of effectiveness data. 1142

The methods used to derive ICPs from effectiveness data have been very variable. In addition to 1143

the factors that may affect the relevance of ICPs derived from efficacy trials, estimates drawn 1144

from effectiveness data may in part reflect the type of immunization program in place and the 1145

extent to which protection of individuals relies on herd immunity rather than the initial and 1146


persisting immune response in the individual. The wider applicability of ICPs derived from such 1147

trials should be viewed in light of how and in what setting the estimates were obtained. 1148

1149

If it is not possible to derive a clinical ICP the interpretation of the human immune response data 1150

may take into account what is known about immunological parameters that correlate with 1151

protection in relevant animal models and any nonclinical ICPs that have been identified (e.g. 1152

from trials that assess passive protection and active immunization). This approach may be the 1153

only option available for interpreting immune responses to some new candidate vaccines. 1154

Nevertheless, ICPs derived wholly from nonclinical data should be viewed with caution and 1155

attempts should be made to obtain a clinical ICP whenever the opportunity arises (e.g. when the 1156

vaccine is used in an outbreak situation). 1157

1158

If conducted, human challenge trials may also provide preliminary evidence supporting an ICP. 1159

Nevertheless, these trials are usually conducted in non-immune healthy adults who are 1160

challenged with organisms that are not identical to, and do not behave like, virulent wild-types. 1161

Therefore these trials may point to a correlation between a specific immunological parameter and 1162

protection, which can be further investigated during the clinical development program. 1163

1164

5.5 Immunogenicity trials 1165

1166

5.5.1 Objectives 1167

1168

The objectives of pre-licensure and post-licensure clinical immunogenicity trials include (but are 1169

not limited to): 1170

i) To select vaccine formulations and posologies (including primary and booster doses) 1171

ii) To bridge the efficacy demonstrated in a specific population and using one vaccine 1172

formulation and posology to 1173

a) The same vaccine when used in other settings or with alternative posologies or 1174

b) A different vaccine intended to protect against the same infectious disease(s) 1175

as a licensed vaccine for which efficacy has been established 1176

iii) To achieve the objectives as in ii) but in the absence of prior efficacy data to which a 1177


bridge can be made 1178

iv) To support co-administration with other vaccines 1179

v) To support maternal immunization with the primary intent to protect the infant 1180

vi) To support major changes to the manufacturing process 1181

vii) To assess lot to lot consistency (8) 1182

1183

Subsections 5.5.2 and 5.5.3 address some general considerations for the selection of endpoints, 1184

the design of comparative immunogenicity trials and the analysis and interpretation of the 1185

results. Subsection 5.6 provides additional details of issues to take into consideration when 1186

designing, analyzing and interpreting comparative immunogenicity trials that have one or more 1187

of objectives i) to vii). 1188

1189

5.5.2 General considerations for trial designs 1190

1191

Immunogenicity trials are almost without exception comparative trials. Comparative trials 1192

include those in which all subjects receive the same vaccine formulation but there are differences 1193

between groups in how or to whom the vaccine is administered (e.g. using a different dose or 1194

dose interval, administering the vaccine to different age groups) and trials in which at least one 1195

of the trial groups receives an alternative treatment, which may be placebo and/or another 1196

licensed vaccine. 1197

1198

The design of comparative immunogenicity trials is driven by the characteristics of the vaccine, 1199

the trial objectives, the stage of clinical development, the trial population, the availability and 1200

acceptability of suitable comparators and what is known about immune parameters that correlate 1201

with protection (including whether or not there is an established ICP). 1202

1203

In comparative immunogenicity trials subjects should be randomized to one of the trial groups at 1204

enrolment. This also applies to trials that enroll sequential cohorts of subjects (e.g. in ascending 1205

dose trials in which at least some subjects are assigned to receive placebo or another vaccine). In 1206

some cases it may be appropriate that subjects who meet certain criteria (e.g. completed all 1207

assigned doses in the initial part of the trial) are re-randomized at a later stage of the trial to 1208


receive a further dose of a test or control treatment. 1209

1210

Whenever possible, comparative immunogenicity trials should be double blind. If the vaccines to 1211

be compared are visually distinguishable, it is preferable that designated persons at each trial site 1212

administer the products. Vaccinees (or their parents/guardians) and all other trial staff should 1213

remain unaware of the treatment assignment. If this is not feasible, or if the vaccines to be 1214

compared are given by different routes or at different schedules, the assays should be conducted 1215

by laboratory staff unaware of the treatment assignment. 1216

1217

In trials intended to provide only descriptive analyses of the immunogenicity data the trial 1218

sample size is usually based on considerations of feasibility and collection of sufficient safety 1219

data to support the design of sequential trials. Trials that aim to assess superiority or non-1220

inferiority between vaccine groups should be sized according to the intended power and the pre-1221

defined margins. 1222

1223

5.5.2.1 Endpoints 1224

1225

The choice of the primary trial endpoint and the range of other endpoints for immunogenicity 1226

trials should take into account Subsections 5.2, 5.3 and 5.4. Protocols should pre-define the 1227

primary, secondary and any other (which may be designated tertiary or exploratory) endpoints. 1228

Trial protocols may pre-define multiple co-primary endpoints: 1229

For vaccines intended to protect against multiple subtypes of the same micro-organism (e.g. 1230

human papillomavirus vaccines, pneumococcal conjugate vaccines) 1231

For combination vaccines, including vaccines that contain multiple micro-organisms (such 1232

as measles, mumps, rubella vaccine) or multiple antigens (such as combination vaccines 1233

used for the primary immunization series in infants) 1234

1235

The following should be taken into consideration when selecting the primary endpoint(s) 1236

following primary vaccination: 1237

1238

i. When an ICP has been established the primary endpoint is usually the percentage of 1239


vaccinees that achieves an antibody level at or above the ICP, which is sometimes referred 1240

to as the seroprotection rate. 1241

1242

ii. When there is no established ICP the primary endpoint is usually based on the parameter 1243

that is known or could be anticipated to best correlate with efficacy (e.g. a measure of 1244

functional antibody or, if no functional assay is available, a measure of total IgG). 1245

In some instances there may not be an ICP but there may be evidence to support 1246

application of a threshold value (i.e. the primary endpoint may be the percentage of 1247

vaccinees that achieves antibody levels at or above the threshold value, which is 1248

sometimes referred to as the responder rate). 1249

If there is no ICP or threshold that could be applied it may be appropriate that the primary 1250

endpoint is based on the seroconversion rate or on some other definition of the magnitude 1251

of the immune response that differentiates responders from non-responders. Comparisons 1252

of post-vaccination seropositivity rates may also be informative if pre-vaccination rates 1253

are very low. 1254

1255

For assessment of the immune response following administration of a vaccine to subjects who 1256

are already primed against one or more micro-organisms or antigens in the vaccine an 1257

anamnestic immune response is anticipated so that seroprotection, seroconversion (when defined 1258

by fold-rise from pre- to post-boost) and seropositivity rates after the booster dose will likely be 1259

very high. In these cases the most sensitive immunological parameter for detecting differences 1260

between groups may be the geometric mean concentration or titre. 1261

1262

After primary vaccination and after any additional doses the results of all immunological 1263

parameters measured should be reported, including seroprotection (if defined), seropositivity and 1264

seroconversion rates, geometric mean concentrations or titres and the reverse cumulative 1265

distributions, regardless of the pre-defined primary endpoint. 1266

1267

5.5.2.2 Exploratory trials 1268

1269

In the initial stages of vaccine clinical development, and when commencing further vaccine 1270


development to substantially modify the initial prescribing information, exploratory trials are 1271

commonly conducted to provide preliminary data on safety and immunogenicity. The assessment 1272

of the immune response may be designated as co-primary with safety or secondary. Exploratory 1273

trials are not usually powered or designed to address specific hypotheses. To obtain a clear 1274

picture of safety, these trials may include a placebo group if this is considered to be acceptable 1275

(e.g. a placebo group is commonly used in initial trials with a new candidate vaccine in healthy 1276

adults). 1277

1278

5.5.2.3 Superiority trials 1279

1280

Trials intended to detect superiority of immune responses are most often conducted during the 1281

selection of candidate vaccine formulations and posologies for further clinical investigation. It is 1282

common that these trials plan to assess whether a specific candidate vaccine formulation elicits 1283

superior immune responses compared to no vaccination against the disease to be prevented 1284

and/or compared to alternative formulations of the candidate vaccine. Initial dose selection trials 1285

are not usually formally powered to demonstrate superiority but this may be considered for 1286

larger trials that are intended to select a final formulation and posology for further investigation. 1287

1288

Superiority trials are also conducted when an adjuvant is proposed for inclusion in the vaccine, 1289

in which case it is usually expected that the immune response to at least one of the antigenic 1290

components of an adjuvanted formulation should be superior to that for a non-adjuvanted 1291

formulation that is otherwise identical. However, if addition of an adjuvant is intended to reduce 1292

the amount(s) of antigen(s) required (which may increase vaccine production capacity) it may 1293

suffice that the adjuvanted formulation with the reduced antigen dose is shown to be at least as 1294

immunogenic (i.e. non-inferior) as a non-adjuvanted formulation containing a higher dose. 1295

1296

Some trials may be designed to assess superiority between certain groups and non-inferiority 1297

between others or to assess superiority of immune responses to single or multiple antigenic 1298

components. For example, whilst adding an adjuvant may improve the immune responses to one 1299

or more antigenic components it should also not have a negative effect that is of potential clinical 1300

significance on the immune responses to all other antigenic components. In addition, a trial may 1301


be designed to establish that specific immune responses are at least non-inferior between trial 1302

groups and, if the pre-defined non-inferiority criteria are met, to then assess whether the 1303

responses are superior. 1304

1305

5.5.2.4 Non-inferiority trials 1306

1307

Most comparative immunogenicity trials are intended to show that the test vaccinated groups 1308

achieve comparable immune responses to the selected reference groups. Not all such trials need 1309

to be formally designed and powered to demonstrate non-inferiority but trials that are intended to 1310

be pivotal (i.e. the application for licensure or to modify the license is to be based mainly or 1311

wholly on the trial) should be adequately designed and powered to demonstrate non-inferiority 1312

using a pre-defined and justifiable non-inferiority margin. It is recommended that protocols and 1313

statistical analysis plans for each trial are developed in conjunction with an appropriately 1314

experienced statistician. 1315

1316

Factors to consider regarding the stringency of the non-inferiority margin include the clinical 1317

relevance of the endpoint, seriousness of the disease to be prevented and the vulnerability of the 1318

target population. More stringent margins may be appropriate when the vaccine is intended to 1319

prevent severe or life-threatening diseases and will be used in particularly vulnerable populations 1320

(e.g. infants and pregnant women). If a new candidate vaccine is known to offer substantial 1321

benefits in terms of safety or improved coverage, less stringent margins may be considered. In 1322

contrast, a more stringent margin could be considered when there is a potential for a downward 1323

drift in immunogenicity such as that which could occur when a new candidate vaccine can be 1324

compared only with vaccines that were themselves approved based on non-inferiority trials (see 1325

Subsection 5.6.2.1). As a result of these considerations it is possible that different non-inferiority 1326

margins may be considered appropriate to interpret immune responses to any one specific 1327

antigenic component in different settings. 1328

1329

As a general rule, for the purposes of establishing non-inferiority between vaccine groups 1330

based on GMT or GMC ratios for antibody titres or concentrations, it is suggested that the 1331

lower bound of the 95% confidence interval around the ratio (test vs. reference vaccine) should 1332


not fall below 0.67. Under certain circumstances, NRAs may consider allowing a lower bound 1333

of 0.5. The criterion should be selected taking into account whether or not an ICP has been 1334

identified. In addition, any marked separations between the reverse cumulative distributions of 1335

antibody titres or concentrations should be discussed in terms of the potential clinical 1336

implications, even if these occur only at the lower or upper ends of the curves. 1337

1338

When comparing seroprotection rates, seroconversion rates or percentages of vaccines with 1339

immune responses that are above a pre-defined threshold, sponsors frequently select a non-1340

inferiority margin of 10%, which gives modest sample sizes. There is very rarely any 1341

justification provided for this margin nor is there any discussion of the possible consequences 1342

of a candidate vaccine eliciting seroprotection or seroconversion rates or percentages with 1343

responses above a pre-defined threshold that are lower those in the licensed vaccine group to 1344

such an extent that the lower 95% confidence interval around the difference (test – reference) 1345

approaches -10%. If a sponsor does pre-define such a margin without adequate justification, 1346

the implications of the actual 95% confidence intervals that are observed should be reviewed 1347

in light of the considerations described above. 1348

1349

5.5.3 Analysis and interpretation 1350

1351

A statistical analysis plan should be finalized before closing the trial database and unblinding 1352

treatment assignments (if these were blinded). This should include any planned interim analyses, 1353

which should be adequately addressed in terms of purpose, timing and any statistical adjustments 1354

required. 1355

1356

The immunogenicity data from all subjects with at least one result for any immunological 1357

parameter measured in the trial should be included in the clinical trial report. The analysis of the 1358

immune response based on any one parameter is commonly restricted to all subjects with a pre-1359

vaccination measurement (if this is to be obtained from all subjects) and at least one post-1360

vaccination measurement. Protocols may also restrict the primary analysis population to subjects 1361

with pre- and post-vaccination results who received all the assigned doses within pre-defined 1362

windows around the intended schedule and had no other major protocol violations (e.g. met the 1363


inclusion and exclusion criteria). Other analysis populations of interest may be pre-defined in 1364

accordance with the primary or secondary objectives (e.g. age sub-groups, pre-vaccination 1365

serostatus). Whatever the pre-defined primary analysis population, all available immunogenicity 1366

data should be presented in the clinical trial report. 1367

1368

If a trial fails to meet the pre-defined criteria for superiority and/or non-inferiority with respect to 1369

any of the antigenic components the possible reasons for the result and the clinical implications 1370

should be carefully considered before proceeding with clinical development or licensure. The 1371

considerations may take into account the basis for setting the pre-defined criteria (e.g. does 1372

failure to meet the criteria strongly imply that lower efficacy may result), the comparisons made 1373

for all other immune parameters measured (e.g. were criteria not met for only one or a few of 1374

many antigenic components of the vaccine), any differences in composition between the test and 1375

the comparator vaccines that could explain the result, the severity of the disease(s) to be 1376

prevented and the overall anticipated benefits of vaccine, including its safety profile. Subsection 1377

5.6 provides some further examples and issues to consider. 1378

1379

Additional analyses of the data that were not pre-specified in the protocol and/or the statistical 1380

analysis plan (i.e. post hoc analyses) should generally be avoided. If conducted, they should 1381

usually be viewed with caution although the results may stimulate further clinical trials to 1382

investigate specific issues. 1383

1384

5.6 Specific considerations for trial design and interpretation 1385

1386

This Subsection should be read in conjunction with Subsection 5.5 1387

1388

5.6.1 Selection of formulation and posology 1389

1390

The vaccine formulation is determined by the numbers of micro-organisms or amounts of 1391

antigens and, if applicable, adjuvant that is to be delivered in each dose as well as the route of 1392

administration. 1393

1394


The vaccine posology for a specific route of administration includes: 1395

Dose content (as for formulation) and volume delivered per dose 1396

Dose regimen (number of doses to be given in the primary series and, if applicable, after the 1397

primary series) 1398

Dose schedule (dose intervals within the primary series and between the primary series and 1399

any further doses) 1400

1401

The vaccine posology for any one vaccine may vary between target populations (e.g. age groups 1402

and according to prior vaccination history) in one or more aspects (content, regimen or 1403

schedule). 1404

1405

The following sections outline the immunogenicity data that are usually generated to support the 1406

vaccine formulation and posology and to assess the need for, and immune response to, additional 1407

doses of the vaccine after completion of the primary series. Section 7 addresses the importance 1408

of the safety profile when selecting vaccine formulations and posologies. 1409

1410

5.6.1.1 Selecting the formulation and posology for initial licensure 1411

1412

The vaccine formulation and posology that is initially approved should be supported by safety 1413

and immunogenicity data, with or without efficacy data, collected throughout the pre-licensure 1414

clinical development programme. At the time of initial licensure the data should at least support 1415

the formulation and posology for the primary series, which may consist of one or more doses. 1416

1417

Depending on the intended formulation of the new candidate vaccine the following 1418

considerations may apply: 1419

1420

i) Whenever a new candidate vaccine contains any micro-organisms or antigens not previously 1421

used in human vaccines, with or without others already used in human vaccines, the initial trials 1422

usually explore the immune responses to different amounts of each of the new micro-organisms 1423

or antigens when given alone in non-immune healthy adult subjects. These trials should describe 1424

the dose-response curve and may indicate a plateau for the immune responses above a certain 1425


dose level. The next trials usually evaluate immune responses to further doses at various dose 1426

intervals to evaluate the kinetics of the immune response as well as any increment in immune 1427

response that is achieved by further doses. The transition from trials in healthy adults to trials in 1428

subjects in the target age range at the time of initial licensure (if this is not confined to young 1429

adults) should occur as soon as this can be supported taking into account the safety profile. 1430

1431

However, evaluating the immune response to each of the new micro-organisms or antigens alone 1432

may not be a feasible undertaking. For example, if the vaccine construct is manufactured in such 1433

a way that production of individual antigens is not feasible then the evaluation of the appropriate 1434

vaccine dose may be based solely on studies with the entire construct. Another example concerns 1435

vaccines intended to protect against multiple subtypes of an organism. In this case, the use of 1436

micro-organisms or antigens that could be regarded as broadly representative in the first trials 1437

may provide some idea of the likely response to other subtypes. Further trials may then explore 1438

formulations that contain increasing numbers of the subtypes with the objective of assessing the 1439

effect of combining them into a single product on the immune response. 1440

1441

ii) For new candidate vaccines that contain known antigenic components not previously 1442

combined together into a single vaccine the initial trials are usually conducted in subjects within 1443

the age ranges approved for licensed vaccines that contain some or all of the same antigenic 1444

components. The aim is to demonstrate non-inferiority of immune responses to each of the 1445

intended antigenic components when combined into a candidate formulation with co-1446

administration of licensed vaccines that together provide all of the same antigenic components. 1447

The same approach applies whenever the antigenic components are not combined into a single 1448

formulation but the contents of more than one pre-formulated product have to be mixed 1449

immediately before administration to avoid a detrimental physico-chemical interaction. 1450

1451

iii) For new candidate vaccines that contain known and one or more new antigenic components 1452

the initial trials may aim to demonstrate non-inferiority of immune responses to each of the 1453

known antigenic components when combined into a candidate formulation with separate 1454

administrations of known and new antigenic components. It may also be informative to include a 1455

control group that receives co-administration of known and new antigenic components. The 1456


exact design depends on the availability of a single licensed vaccine containing the known 1457

antigenic components or whether more than one licensed vaccine has to be given. 1458

1459

iv) For any vaccine formulation to which an adjuvant is to be added there should be adequate 1460

data already available (which may apply to known adjuvants) or data should be generated (new 1461

adjuvants or when using any adjuvant with a new antigenic component) to demonstrate that 1462

addition of the adjuvant elicits a superior immune response to one or more antigenic components 1463

without a potentially detrimental effect on any other antigenic components. Alternatively, data 1464

should demonstrate that including the adjuvant allows for the use of a much lower dose of an 1465

antigenic component to achieve the desired level of immune response. Trials should evaluate a 1466

sufficient range of combinations of antigenic components and adjuvant to support the final 1467

selected formulation. 1468

1469

v) The total data generated should be explored to identify the criteria to be applied for the 1470

determination of an appropriate shelf-life of the vaccine. This is usually of particular importance 1471

to vaccines that contain live micro-organisms. Depending on data already generated, it may be 1472

necessary to conduct additional trials with formulations known to contain a range of micro-1473

organism numbers or antigen doses to identify appropriate limits at end of shelf-life. 1474

1475

vi) Comparative immunogenicity trials may be needed to determine schedules appropriate for 1476

specific target populations, taking into account the urgency to achieve protective immunity (i.e. 1477

based on diseases to be prevented and their epidemiology). The data generated across all the 1478

trials should determine the minimum period that should elapse between doses and the effects of 1479

delaying doses to support acceptable windows around scheduled doses. Additionally, for some 1480

vaccines it may be useful to explore the shortest time frame within which doses may be 1481

completed without a detrimental effect on the final immune response (e.g. for vaccines for 1482

travelers who may need to depart at short notice and for vaccines intended to provide post-1483

exposure prophylaxis). 1484

1485

The assessment of the effects of dose interval and the total time taken to complete the primary 1486

series is a particular issue for vaccines intended for use in infants due to the very wide range of 1487


schedules in use in different countries (e.g. 3-dose schedules include 6-10-14 weeks and 2-4-6 1488

months). In general, experience indicates that the magnitude of the post-primary series immune 1489

responses broadly correlates with the age of infants at the time of the final dose. If a trial using a 1490

6-10-14 weeks or 2-3-4 months schedule demonstrates highly satisfactory immune responses it is 1491

reasonable to expect that schedules that either commence later in infancy, use longer dose 1492

intervals and/or in which the final dose is given at 5-6 months or later will also be highly 1493

satisfactory. In contrast, the results of the latter types of schedules cannot be used to support use 1494

of earlier and more condensed schedules. 1495

1496

vii) All of the data generated in accordance with points i) to vi) should be taken into account 1497

when selecting the final formulation and posology or posologies. The selection process is more 1498

straightforward if there are established ICPs that can be applied to interpretation of the results for 1499

at least some of the antigenic components. In the absence of an ICP, which frequently applies to 1500

new micro-organisms or antigens, the posology may be selected from considerations of any 1501

plateau effects that are observed and the safety profile of various doses and regimens. 1502

1503

It is not unusual that the final selected formulation and posology to some extent represents a 1504

compromise between immunogenicity and safety or, for combination vaccines, between the 1505

potential benefits of a vaccine that can protect against multiple types of infectious disease with 1506

some negative effects on immune response that may occur. These negative effects may result 1507

from a physicochemical interaction between vaccine components and/or a negative immune 1508

interference effect for some antigenic components with or without a positive immune 1509

interference effect for some others. The rationale for the final selection requires careful 1510

discussion in the application dossier. 1511

1512

5.6.1.2 Amending or adding posologies after initial licensure 1513

1514

Clinical trials conducted after first licensure may be designed to address one or more of the 1515

following: 1516

a. Change the number of doses or dose intervals. In this case the control group should be 1517

vaccinated using the licensed posology and the trial should be conducted in a population 1518


for which the vaccine is already licensed. 1519

b. Use of the licensed posology in a new population (e.g. in subjects who are younger or 1520

older than the currently licensed age group; in subjects with specific underlying 1521

conditions, such as immunosuppression). In this case the trial should compare use of 1522

the licensed posology in the new target population and the population for which the 1523

vaccine is already licensed. 1524

c. Use of an alternative to the licensed posology in a new population. In this case the 1525

alternative posology administered to the new population should be directly compared 1526

with the licensed posology in the licensed population. 1527

d. Support alternative routes of administration for the licensed formulation (e.g. adding sub-1528

cutaneous or intra-dermal injection to intra-muscular use). 1529

1530

Post-licensure clinical trials may also be conducted to support changes in formulation. 1531

Formulation changes other than adding or removing a preservative or removing thiomersal from 1532

the manufacturing process usually result in a modified product that is considered to be a new 1533

candidate vaccine from a regulatory standpoint (i.e. it would require a new application dossier 1534

and adequate trials to support separate licensure). 1535

1536

5.6.1.3 Post-primary doses 1537

1538

a. Need for post-primary doses 1539

1540

The need to administer additional doses, and the timing of these doses, may be determined 1541

before and/or after first licensure. 1542

1543

To date, very few licensed vaccines are recommended only for use in a primary series. Examples 1544

include inactivated hepatitis A vaccines and hepatitis B vaccines containing recombinant surface 1545

antigen [HBsAg] for which very long term follow-up continues to suggest that additional doses 1546

are not necessary to maintain protection in those who had a robust immune response to the 1547

primary series. For all other vaccines one or more additional doses of the same or another 1548

vaccine that protects against the same disease(s) is recommended or the prescribing information 1549


states that it is not yet known whether further doses will be necessary. 1550

1551

If experience with other similar vaccines clearly indicate that additional doses of a new candidate 1552

vaccine will be needed the clinical development program should incorporate this in the overall 1553

assessment of immune responses. 1554

1555

If it is not known whether post-primary doses of a new candidate vaccine will be needed to 1556

maintain protection it is preferable that this should be determined from long-term follow-up of 1557

subjects who were enrolled in efficacy trials and/or from post-licensure effectiveness trials. 1558

Although the long-term monitoring of antibody persistence is important, these data alone cannot 1559

determine if another dose is needed unless there is evidence or a strong reason to expect that 1560

failure to maintain circulating antibody above a certain level (e.g. above the ICP if there is one) 1561

is associated with risk of breakthrough disease (even when the primary series of the vaccine 1562

elicited an immune memory response). 1563

1564

Until it is clear whether or not additional doses are needed, it is prudent to plan to obtain data on 1565

the immune response to additional doses at different intervals after the last dose of the primary 1566

series so that data are available should it become clear that an additional dose is required. 1567

1568

b. Assessment of priming during the primary series 1569

1570

Not all vaccines elicit a T-cell-dependent immune response that results in priming of the immune 1571

system and an anamnestic response to further doses. The administration of post-primary doses of 1572

a new candidate vaccine that contains one or more micro-organisms or antigens not previously 1573

used in human vaccines provides an opportunity to assess whether there was successful priming 1574

of the immune system during the primary series, in which case subsequent doses will serve to 1575

boost the immune response (see Subsection 5.2). 1576

1577

When assessing the immune response to additional doses and determining whether or not the 1578

primary series elicited immune memory the following should be taken into account: 1579

a. Trials in which additional doses are administered may be extension phases of primary series 1580


trials or new trials in subjects with documented vaccine histories. 1581

b. When assessing whether the primary series elicited immune memory the optimal design is to 1582

compare subjects who previously completed a full primary series of the candidate vaccine 1583

with a control group consisting of subjects not previously vaccinated. Control subjects 1584

should be matched for age and for any host or demographic factors that might impact on 1585

their immune response (e.g. they should be resident in similar areas so that any natural 1586

exposure is likely similar). 1587

c. If the new candidate vaccine elicited immune memory in the primary series the immune 1588

response to the additional (i.e. booster) dose should usually be superior to that observed in 1589

individuals who have not been vaccinated against the disease to be prevented based on 1590

comparisons of the geometric mean concentrations or titres of antibody. The percentages 1591

that achieve seropositivity or seroprotection (as defined) may not be different between the 1592

two groups if a single dose of the vaccine is highly immunogenic even in unprimed 1593

individuals. 1594

d. The immune response to the additional dose in primed and unprimed subjects may also be 1595

differentiated based on the rapidity of the rise in antibody levels (faster in primed) and in 1596

terms of antibody avidity (greater in primed). 1597

e. If the immune response as measured by geometric mean antibody concentrations or titres in 1598

the primed group is not superior to that in controls this does not always mean that the 1599

primary series did not elicit immune memory. For example, this may occur when natural 1600

priming has occurred in a substantial proportion in the control group that was not previously 1601

vaccinated against the disease to be prevented, in which case the rapidity of response and 1602

measurements of avidity may also not be distinguishable between groups. If natural priming 1603

has occurred it may or may not be detectable from pre-vaccination antibody levels in the 1604

control group. 1605

f. If an immune memory response is elicited in the primary series it may be possible to achieve 1606

a robust anamnestic response using a much lower dose of an antigenic component compared 1607

to the primary series. A lower boosting dose may also provide a better safety profile (e.g. as 1608

occurs with diphtheria toxoid). 1609

g. For polysaccharide-protein conjugate vaccines that elicit immune memory it may be 1610

informative to compare boosting with the same type of conjugate used for priming with an 1611


alternative conjugate (e.g. to prime with a tetanus toxoid conjugate and boost with a 1612

CRM197 conjugate and vice versa). 1613

h. It may also be informative to assess the ability of a candidate vaccine to achieve cross-1614

priming by using heterologous antigenic components for priming and boosting. This may be 1615

assessed by comparing boosting with the same vaccine used to prime with administration of 1616

a formulation (which may be a licensed vaccine or an unlicensed product manufactured 1617

specifically for the trial) containing a different micro-organism or antigen that is known to 1618

be closely related but not identical to that in the vaccine (e.g. material derived from an 1619

influenza virus of a different clade). 1620

i. Elicitation of an immune memory response to a vector for an antigen after the first dose(s) 1621

may interfere with or wholly prevent the immune response to the antigen after subsequent 1622

doses (e.g. this may be observed when using adenoviruses capable of infecting humans as 1623

live viral vectors). It is essential to understand whether or not this occurs since it may 1624

necessitate the use of a different vector for the antigen or an entirely different vaccine 1625

construct to deliver subsequent doses. 1626

j. There are some antigens that not only do not elicit an immune memory response but also 1627

demonstrate hypo-responsiveness to further doses. The best known examples are some of 1628

the unconjugated meningococcal and pneumococcal polysaccharides (17, 18). In the past 1629

these were sometimes administered to assess whether corresponding conjugated 1630

polysaccharides had elicited immune memory in the primary series based on the premise 1631

that this would better mimic the immune response to natural exposure compared to 1632

administration of a further dose of the conjugate. This practice is not recommended since it 1633

is possible that a dose of unconjugated polysaccharide could result in blunted immune 1634

responses to further doses of the conjugate. 1635

1636

5.6.2 Using immunogenicity data to predict efficacy 1637

1638

Immunogenicity data may be used to predict efficacy with varying levels of confidence when: 1639

a. There is a well-established ICP that can be used to interpret the immune responses to a 1640

specific antigenic component (see Subsections 5.4 and 5.5). Comparative 1641

immunogenicity trials are recommended since they provide a control for interpretation of 1642


any unexpected findings and for safety. Depending on the objectives the comparator may 1643

be the same vaccine used as currently licensed or a licensed vaccine that has been widely 1644

used with no known problems regarding its effectiveness and which contains all or as 1645

many as possible of the same antigenic components as the candidate vaccine. 1646

b. It is possible to use immune responses to bridge to estimates of vaccine efficacy obtained 1647

from well-designed clinical trials (i.e. to conduct bridging trials); see Subsection 5.6.2.1. 1648

c. There is no ICP nor is it possible to bridge to a prior demonstration of efficacy; see 1649

Subsection 5.6.2.2. 1650

1651

5.6.2.1 Bridging to efficacy data 1652

1653

There are two main situations to consider. In both cases comparative immunogenicity trials 1654

designed to demonstrate non-inferiority are recommended. The choice of comparator is a critical 1655

factor for interpretation of the results. 1656

1657

i) Modifying the use of the same vaccine for which efficacy has been estimated 1658

1659

As described in Section 6, vaccine efficacy trials are usually conducted in specific target 1660

populations, characterised by factors such as age, region (which may define endemicity for some 1661

infectious diseases) and health status, using the intended final vaccine posology. Before or after 1662

initial licensure trials may be conducted with the aim of extending the use of the vaccine to other 1663

populations and/or to support alternative posologies. 1664

1665

When a different age group or posology is proposed or when extending use from 1666

immunocompetent to immunocompromised subjects it is usually very clear that a bridging trial is 1667

necessary. Whether or not a bridging trial is necessary to support use in regions other than where 1668

the estimate of efficacy was obtained requires careful consideration. Such trials should be 1669

required for licensure only if there are compelling scientific reasons to expect that the immune 1670

response to the vaccine, and therefore its efficacy, could be significantly different due to host 1671

factors (such as common underlying conditions that may affect immune responses) and/or 1672

geographical factors (such as distributions of subtypes of organisms, levels of natural exposure 1673


and for trials in infants the possibility that high levels of maternal antibody could interfere with 1674

responses to the primary series). 1675

1676

The usual trial design involves a direct comparison between the new population and/or posology 1677

and a control group in which subjects representative of the efficacy trial population receive the 1678

previously studied posology. It may also be acceptable that an indirect comparison is made with 1679

the immunogenicity data that were obtained during the efficacy trial, in which case the vaccine 1680

formulation and assay used should be the same as used in the efficacy trial whenever possible. 1681

1682

a. If the vaccine used in the efficacy trial is no longer available the comparator should be as 1683

similar as possible to the original. Over time, it may be that the only bridge back to the 1684

efficacy data is via a comparison with a licensed vaccine that was itself licensed based on a 1685

bridging efficacy trial. As the number of bridging steps that has occurred between the 1686

original efficacy data and the licensed comparator vaccine increases, so the reliance that may 1687

be placed on a demonstration of non-inferiority to predict efficacy is weakened. This 1688

consideration also applies when the vaccine for which efficacy was estimated has been 1689

extended based on bridging efficacy for the shared subtypes (e.g. when additional subtypes 1690

have been added) and the extended vaccine has replaced the original vaccine in the market. 1691

b. If the assay has changed and has not been or cannot be directly compared to the original 1692

assay used during the efficacy trial it may be possible to re-assay stored sera collected 1693

during the prior efficacy trial in parallel with the sera from the new trial population. 1694

1695

If it remains unknown which immunological parameter best correlates with efficacy it is 1696

preferable that the primary comparison between vaccines is based on functional antibody 1697

whenever this is feasible. 1698

1699

ii) Inferring the efficacy of a new candidate vaccine 1700

1701

In this case the main evidence of efficacy for licensure comes from one or more bridging 1702

efficacy trials. The same considerations regarding primary comparison, choice of comparative 1703

vaccine and assay apply as described above. 1704


1705

If the new candidate vaccine is an extended version of a licensed vaccine and/or it contains 1706

additional subtypes of an organism not included in a licensed vaccine the interpretation of the 1707

immune responses to the unshared types in a comparative immunogenicity trial is not 1708

straightforward. Approaches that could be considered include comparing immune responses to 1709

each additional subtype with a mean response across all subtypes or the lowest response to an 1710

individual subtype included in the vaccine for which efficacy was demonstrated. Both of these 1711

approaches may provide a route to licensure but the limitations of these comparisons to predict 1712

efficacy should be taken into account when considering the overall benefit-risk relationship for 1713

the new vaccine and the collection of effectiveness data in the post-licensure period is 1714

recommended. 1715

1716

5.6.2.2 Other approaches 1717

1718

When there is no ICP nor is it possible to bridge to a prior demonstration of efficacy licensing a 1719

new candidate vaccine is problematical. This situation is most likely to apply to new vaccines 1720

against rare infectious diseases such as some viral haemorrhagic fevers, for which outbreaks do 1721

not occur in substantial numbers of persons or are of short durations, and some micro-organisms 1722

that could be used for bioterrorism purposes. Another important situation is the development of 1723

influenza vaccines against potential pandemic strains. 1724

1725

Approaches may include establishing a nonclinical model of efficacy that is thought to be 1726

relevant to the human infection and identifying which immunological parameter best correlates 1727

with protection (and if possible a putative ICP), trials of natural infection and protection against 1728

further disease and any passive protection data that may be available from nonclinical or clinical 1729

trials. If a vaccine has already been licensed based on evidence derived from one of these 1730

approaches any changes to the vaccine usage is subject to the same issues. 1731

1732

Although licensure of vaccines based on these approaches means that it is not likely to be 1733

possible to achieve a high level of confidence in the level of efficacy in humans, having available 1734

vaccines that have already been subjected to a full review of quality and nonclinical data as well 1735


as at least some safety and immunogenicity data in humans does mean that they could be ready 1736

for rapid use in an emergency situation. Nevertheless, for these products it is particularly 1737

essential that protocols are developed in advance of any such emergency so that adequate data 1738

can be collected to assess efficacy/effectiveness whenever the opportunity arises. 1739

1740

5.6.3 Co-administration trials 1741

1742

Comparative immunogenicity trials intended to support co-administration of a vaccine with one 1743

or more other vaccines (i.e. administration at the same time but using different limbs for 1744

injection or multiple routes of administration) should demonstrate non-inferiority for immune 1745

responses to each of the co-administered antigenic components (see Subsection 5.5.3). The 1746

immunological parameters applied to each comparison may differ depending on vaccine content. 1747

It should be noted that co-administration may also enhance the immune response to certain 1748

antigens but so far there have not been instances in which this has been regarded as a cause for 1749

concern since the safety of co-administration has been acceptable. 1750

1751

When there are multiple licensed products containing the same antigenic components that could 1752

be co-administered with the vaccine under trial (e.g. combination vaccines intended for the 1753

routine infant primary immunization series) it is not feasible nor should it be necessary to 1754

conduct trials with each licensed product. The vaccine(s) chosen for trial should be as 1755

representative as possible of the range of licensed products. 1756

1757

An exception arises when there are several different types of polysaccharide-protein conjugate 1758

vaccines available that may be co-administered with the vaccine under trial. This is usually only 1759

an issue when the vaccine under trial contains protein that is the same as, or similar to, that in 1760

available conjugates. In this case it is important to appreciate that the results obtained with any 1761

one conjugate may not be applicable to other types of conjugate (e.g. lack of immune 1762

interference with a tetanus toxoid conjugate does not rule out that this could occur with a 1763

CRM197 conjugate). 1764

1765

If multiple doses of the co-administered vaccines are needed it is usual that the comparison 1766


between groups is made only after completion of all doses. The schedule at which the vaccines 1767

are co-administered may also be an issue if there are several possible alternatives (e.g. as applies 1768

to vaccines for the primary immunization series in infants and for vaccines against hepatitis A 1769

and B). Consideration may be given to using a schedule that is most likely to detect an effect if 1770

there is one. 1771

1772

These trials usually have the following designs: 1773

Randomized parallel group trials in which different groups of subjects receive the vaccine 1774

under trial alone, the vaccine intended for co-administration and both together. If there is 1775

more than one additional vaccine that may be co-administered at the same time additional 1776

groups should receive each of these vaccines alone. In this case it is useful for interpretation 1777

of any observed effects to also add groups that each receives the vaccine under trial with one 1778

of the additional vaccines as well as a group that receives them all together. 1779

Randomized trials that use a staggered administration design. This approach is necessary 1780

when it is not possible to withhold any antigenic components to be co-administered (e.g. 1781

during the infant primary schedule). In these trials one group receives the co-administered 1782

vaccines at a chosen schedule while the control group receives either the vaccine under trial 1783

or the vaccine to be co-administered at the same schedule as the test group and the other 1784

vaccine is given one month later (or other appropriate interval). For completeness, an 1785

additional control group may be used in which the order of staggered vaccine 1786

administrations is reversed. The final dose and sampling occurs at least one month later 1787

compared to the co-administration group which, in infants, could have some impact on the 1788

magnitude of the immune response. 1789

1790

5.6.4 Immunization of pregnant women 1791

1792

5.6.4.1 Aims of immunization during pregnancy 1793

1794

Immunization during pregnancy may be undertaken with the primary aim to: 1795

1796

a. Protect the mother. For any candidate vaccine under development for prevention of an 1797


infectious disease in which the target population includes adolescents and adults there 1798

is a need to consider the importance of generating data in pregnant women to support 1799

its use. The considerations should take into account the nature of the vaccine construct 1800

(e.g. does the vaccine contain a live organism that is replication-competent), whether 1801

pregnant women can reasonably avoid exposure to an infectious agent (e.g. by not 1802

travelling) and whether they may have the same risk of exposure but a greater risk of 1803

experiencing severe disease compared to non-pregnant women of the same age. 1804

1805

b. Protect the infant from an infectious disease for a limited period after birth by means of 1806

trans-placental transfer of maternal antibody. In this case there may be a potential benefit 1807

to the mother (e.g. influenza, acellular pertussis) or no or negligible potential benefit to 1808

the mother (e.g. respiratory syncytial virus and Streptococcus Group B). 1809

1810

5.6.4.2 Dose-finding in pregnancy 1811

1812

For new candidate vaccines intended for use in pregnant women and for licensed vaccines not 1813

authorized for use in pregnancy the first clinical trials to support this use should be conducted in 1814

non-pregnant adults, including or consisting only of women of child-bearing age (19). Once 1815

there are adequate relevant nonclinical data with satisfactory findings and some data on immune 1816

responses in non-pregnant women data should be obtained from pregnant women, covering a 1817

representative age range, so that the effects of pregnancy on the immune response can be 1818

evaluated. The doses tested initially in pregnant women should be based on the non-pregnant 1819

adult data but may need to be adjusted (in terms of antigen dose or dose regimen) after review of 1820

results from initial trials due to the effects of pregnancy on the immune system. Additional 1821

considerations for dose-finding when the aim is primarily to protect the infant are provided in 1822

Subsection 5.6.4.3. 1823

1824

In all trials conducted in pregnant women adequate mechanisms should be in place to document 1825

the outcome of the pregnancy, including the duration of gestation at time of delivery, the 1826

condition of the infant at birth and the presence of any congenital conditions. Depending on the 1827

type of vaccine, it may also be considered appropriate to collect information on developmental 1828


milestones at least during the first few years of life. 1829

1830

5.6.4.3 Passive protection of infants 1831

1832

Transfer of IgG across the placenta does not occur to any extent until the third trimester. If the 1833

vaccine is not expected to benefit the mother, then administration in the third trimester should be 1834

studied. If the aim is also to provide some benefit to the mother, administration earlier in 1835

pregnancy should be studied. In this case, since the immune response to vaccination changes as 1836

pregnancy progresses and women do not always access healthcare early on, the effect of dosing 1837

at different times during pregnancy should be evaluated. 1838

1839

If it is expected that a substantial proportion of adults are likely to already have evidence of 1840

humoral immunity against the infectious disease to be prevented so that the aim of vaccination 1841

during pregnancy is to increase the amount of antibody transferred to the fetus, the trials in 1842

pregnant women may need to include exploration of doses and, if more than one dose is needed, 1843

dose intervals in seropositive as well as seronegative adults. 1844

1845

When the aim is primarily to protect the infant, dose-finding trials in pregnant women should 1846

include measurement of antibody levels in cord blood samples taken at delivery. The number of 1847

samples obtained should be sufficient to provide an estimate of inter-individual variability. In 1848

addition, efforts should be made to collect cord blood data that cover a range of times between 1849

maternal vaccination and delivery, that allow for evaluation of the effects of unexpected early 1850

delivery and which measure the impact of placental dysfunction (e.g. based on infants of low 1851

birth weight for their gestational age). The cord blood levels in infants born to vaccinated 1852

mothers who receive the final selected vaccine posology should be clearly superior to that in 1853

infants born to mothers who were not vaccinated, regardless of the pre-vaccination serostatus of 1854

the mothers. Secondary analyses could examine whether this finding also applies within subsets 1855

of mothers who were seronegative or seropositive prior to vaccination. 1856

1857

The duration of detectable maternal antibody in infants should be documented. To avoid multiple 1858

bleeds in individual infants this may be documented by randomization of mothers such that their 1859


infants are sampled only once or a few times at staggered defined intervals so that the total data 1860

are used to describe the antibody decay curve. These data are particularly important when it is 1861

planned that passive protection via maternal antibody will be followed by active vaccination of 1862

infants against the same antigen(s). 1863

1864

If there is an immune correlate of protection established for the infectious disease to be 1865

prevented the aim of the immunogenicity trials should be to identify a maternal vaccination 1866

regimen that results in cord blood levels that exceed the ICP in a high proportion of new born 1867

infants. If there is no ICP, an efficacy trial in infants is usually needed (see Section 6). 1868

1869

5.6.5 Changes to the manufacturing process 1870

1871

Changes made to the product composition (e.g. addition of, removal of, or change in adjuvants or 1872

preservatives) or manufacture (changes to process, site or scale) during the pre-licensure clinical 1873

development program or after licensure do not always need to be supported by comparative 1874

clinical immunogenicity trials between the prior and the newer products. 1875

1876

For example, it is common that the scale of manufacture changes during the pre-licensure 1877

development program but this step alone would not be expected to have a clinically significant 1878

effect in the absence of other changes. In addition, the later confirmatory trials usually use 1879

product from final scale process. Also, any clinical effects of changes to the manufacturing 1880

process during the pre-licensure program may be evident from the results of sequential trials in 1881

similar populations or may not matter if the pivotal immunogenicity and/or efficacy trials use 1882

vaccine made using the final process. If this is not the case, and for all changes that are made 1883

post-licensure, consideration must be given to whether a clinical trial to compare vaccine 1884

manufactured using the prior and new processes is required. This decision must be taken on a 1885

case by case basis after a full evaluation of the in-vitro and any nonclinical in-vivo data 1886

describing and supporting the change. It is usually acceptable that a single lot of vaccine made 1887

using each process is sufficient for the comparison. 1888

1889

In the post-licensure period there may be many changes to the manufacturing process over time. 1890


Over time it is possible that each one of these was considered too minor to merit conduct of a 1891

clinical trial but the product that results from multiple minor changes could be substantially 1892

different to that which was initially licensed. When considering the potential impact of what 1893

seems to be a relatively minor change to the production process that, not alone, would merit a 1894

clinical trial it may be important to consider the full history of changes that have been allowed 1895

without clinical data and to consider whether the sum total of changes could have a clinical 1896

impact. In this situation, when many years have passed, a clinical trial of the current compared to 1897

the original licensed vaccine will not be possible. If disease surveillance suggests that there could 1898

be a problem with vaccine effectiveness, a clinical trial that compares the current vaccine with 1899

another licensed vaccine for which there is a lot of clinical experience may be considered useful. 1900

1901

5.6.6 Lot-to-lot consistency trials 1902

1903

Some NRAs request lot-to-lot consistency trials during the pre-licensure clinical development 1904

program for all new candidate vaccines. Where these trials are not requested as a routine they 1905

may be considered for certain types of vaccines where there is inherent variability in 1906

manufacture of the product. If requested, the rationale for conducting the trial and the objectives 1907

should be very clear. 1908

1909

In these trials the usual expectation is that 95% confidence interval around each pairwise 1910

comparison of the post-vaccination geometric mean concentrations/titres falls within pre-defined 1911

limits. The clinical implications of results that show that one or more comparisons do or do not 1912

meet the pre-defined criteria set around the ratios are unknown and interpretation of the results 1913

should take into account all of the available immune response data. 1914

1915


6. Efficacy and effectiveness 1916

1917


Approaches to determination of efficacy 1919

Human challenge trials 1920

Preliminary and confirmatory (pivotal) efficacy trials 1921

Design and conduct of efficacy trials, including control groups 1922

Approaches to determination of vaccine effectiveness 1923

1924

6.1 Approaches to determination of efficacy 1925

1926

6.1.1 Human challenge trials 1927

1928

In some settings it may be useful and appropriate to obtain an initial assessment of vaccine 1929

efficacy from human challenge trials in which vaccinees are deliberately exposed to an infectious 1930

agent in a controlled setting. Human challenge trials are not always feasible or appropriate, as 1931

discussed in Appendix 1. When they can be performed, human challenge trials have potential to 1932

streamline and so accelerate vaccine development. They may be of particular use: 1933

o When there is no appropriate nonclinical model (e.g. when a candidate vaccine is 1934

intended to protect against an infectious disease that is confined to humans). 1935

o When there is no known immunological correlate of protection. 1936

o When vaccine efficacy trials (as described above and in detail in the sections that follow) 1937

are not feasible. 1938

1939

Like all model systems human challenge trials have limitations in terms of their relevance to 1940

natural infection and their ability to predict protection under very variable circumstances (e.g. in 1941

terms of time elapsed between vaccination and exposure to a pathogen and the impact of 1942

pathogen dose on development of clinically apparent infection). Nevertheless, they may suffice 1943

to rule out vaccines or doses that seem unlikely to have useful protective efficacy and to select 1944

the most promising formulations and regimens for further trial. See Appendix 1 for further 1945

information. 1946


1947

Later on in the clinical development program, usually after safety and immunogenicity trials 1948

have identified one or more potentially effective vaccination regimens for further evaluation, 1949

vaccine efficacy may be assessed against naturally acquired infectious disease. 1950

1951

6.1.2 Preliminary efficacy trials 1952

1953

Based on the available safety and immunogenicity data it may be considered appropriate to 1954

evaluate vaccine efficacy initially in dose-finding trials (which may include different doses 1955

and/or different numbers of doses or dose intervals) or in small-scale trials that evaluate a single 1956

vaccination regimen before proceeding to confirmatory (pivotal) trials. 1957

1958

Whenever possible the general features of these trials (such as case definitions and method of 1959

case ascertainment) should resemble those expected to be applied in confirmatory trials of 1960

efficacy. However, it is sometimes the case that preliminary efficacy trials are used to inform the 1961

final design of confirmatory efficacy trials. For example: 1962

o By applying various case definitions the results may be used to identify or refine the most 1963

appropriate case definition for confirmatory trials. 1964

o By exploring efficacy in specific subgroups in preliminary trials the confirmatory trials 1965

may be designed to ensure adequate numbers of cases per subgroup of interest. 1966

o The method of case ascertainment used may be assessed for feasibility in larger trials 1967

with a greater number of, and more geographically widespread, trial sites. 1968

o The immunogenicity and efficacy data may be used to support a provisional assessment 1969

of potential correlates of protection. 1970

1971

If the candidate vaccine is intended to prevent a severe and/or life-threatening infectious disease 1972

for which there is no, or at least no very satisfactory, vaccine already available, individual NRAs 1973

may agree to accept an initial application for licensure based on one or more preliminary efficacy 1974

trial or trials. In these cases it is essential that sponsors and NRAs should discuss and agree the 1975

main features of the design of the trials before initiation, including the sample size, so that, 1976

subject to promising results, the data may be considered robust and sufficient. 1977


1978

The availability of a vaccine licensed on the basis of preliminary efficacy data has potentially 1979

important implications for the acceptability and feasibility of initiating or completing 1980

confirmatory efficacy trials that include a control group that does not receive active vaccination. 1981

These issues should be discussed between NRAs and sponsors so that expectations for provision 1982

of confirmatory efficacy data are agreed prior to the start of any trials that could potentially 1983

support initial licensure. 1984

1985

6.1.3 Confirmatory (pivotal) efficacy trials 1986

1987

A single confirmatory vaccine efficacy trial or more than one trial may be conducted, depending 1988

on considerations described in Subsection 6.2 below. 1989

1990

In pivotal efficacy trials, the primary objective is usually to estimate vaccine efficacy over a pre-1991

defined time frame after completion of the primary vaccination schedule, which may comprise 1992

one or more doses. Confirmatory trials may evaluate a single or more than one vaccination 1993

regimen and may or may not include evaluations of efficacy before and after booster doses. As 1994

applicable to the individual candidate vaccine, a range of secondary efficacy objectives may be 1995

defined although the trial will not be formally powered for these analyses. 1996

1997

6.2 Design and conduct of efficacy trials 1998

1999

The protective efficacy of a vaccine against a specific infectious disease is usually defined as the 2000

reduction in the chance of developing the disease after vaccination relative to the chance when 2001

not vaccinated as determined in a prospective randomized controlled trial. Vaccine efficacy (VE) 2002

is therefore derived from the proportionate reduction in disease attack rate (AR) between the 2003

control group that did not receive vaccination against the infectious disease potentially 2004

preventable by the candidate vaccine (ARU) and the vaccinated (ARV) group(s). VE can be 2005

calculated from the relative risk (RR) of disease among the vaccinated group as (ARU-2006

ARV/ARU) x 100 and (1-RR) x 100. 2007

2008


Much less often, vaccine efficacy may be determined in a prospective randomized trial in which 2009

the efficacy of the candidate vaccine is compared to that of a licensed vaccine intended to 2010

prevent the same infectious disease. 2011

2012

The following sections consider issues that apply to both types of trial, including some specific 2013

trial designs that may be considered along with some issues for analysis of the data. Details of 2014

statistical methodologies are beyond the scope of this guidance and only broad principles are 2015

described. 2016

2017

6.2.1 Selection of trial sites 2018

2019

Vaccine efficacy trials require the presence of a sufficient burden of clinical disease to enable 2020

estimates to be obtained from feasible numbers of subjects and within a reasonable timeframe. 2021

The infectious disease to be prevented may occur at sufficiently high rates to enable efficacy 2022

trials to be conducted only in confined areas. Even when the disease to be prevented is more 2023

widespread, it may be necessary to confine efficacy trials to specific affected areas for reasons 2024

that may include feasibility of dealing with multiple NRAs and ethics committees, need to ensure 2025

adequacy of monitoring and desire to accumulate representative numbers of cases due to specific 2026

serotypes or subtypes. 2027

2028

Sponsors may have to conduct feasibility assessments to accurately ascertain clinical disease 2029

rates in various age subgroups of populations before selecting trial sites. Any nationally-2030

recommended non-vaccine-related preventive measures that are in place (e.g. prophylactic drug 2031

therapy in high risk individuals or settings, use of insect repellents and bed nets) should be 2032

identified and the trial should be conducted against a background of these additional 2033

interventions. 2034

2035

Trial sites need to be sufficiently accessible to allow regular monitoring visits. Sponsors may 2036

have to engage in site capacity building exercises prior to trial initiation, including training of 2037

study personnel, and may need to provide essential infrastructure to support the trial (e.g. to 2038

ensure that there are adequate blood collection and processing facilities, refrigeration facilities 2039


suitable for the vaccine and/or sera, competent laboratories, data handling capacity and 2040

communication methods to allow electronic randomization schemes, rapid reporting of safety 2041

data or other trial issues to the sponsor). 2042

2043

6.2.2 Candidate (test) vaccine group(s) 2044

2045

If previous data do not support selection of a single dose or regimen of the candidate vaccine for 2046

assessment of efficacy, trials may include one or more groups in which subjects receive the 2047

candidate vaccine (e.g. more than one dose or schedule may be evaluated). In some instances one 2048

or more placebo doses may need to be interspersed with candidate vaccine doses to enable 2049

matching of all regimens under trial in a double-blind design (e.g. if 2 or 3 doses of the candidate 2050

vaccine are to be compared with the control group). 2051

2052

6.2.3 Control (reference) group(s) 2053

2054

Control groups comprise all subjects who do not receive the candidate vaccine. Usually only one 2055

control group is enrolled in any one trial. On occasion, it may be considered important to include 2056

more than one of the possible types of control groups that are discussed below. 2057

2058

6.2.3.1 Control groups not vaccinated against the infectious disease to be prevented 2059

2060

In most cases vaccine efficacy trials employ a control group that does not receive vaccination 2061

against the disease to be prevented by the candidate vaccine. In double-blind trials the control 2062

group may receive: 2063

2064

o A true placebo (i.e. material without any pharmacological activity). This has the advantage 2065

of providing safety data against a control that has no pharmacologically active components. 2066

However, the use of an injectable placebo may not be acceptable to one or more of NRAs, 2067

ethics committees, investigators, trial subjects or their parents/guardians at least in some age 2068

groups (e.g. there may be particular objections raised against true placebo injections in 2069


infants). In contrast, there is usually no objection to use of a true placebo when the 2070

candidate vaccine is administered orally or by nasal installation. 2071

2072

o If a true placebo is not acceptable to one or more of the above interested parties the control 2073

group may receive a licensed vaccine that has no effect on the infectious disease to be 2074

prevented by the candidate vaccine but may have some benefit for recipients. In some cases 2075

both licensed vaccine and placebo doses may have to be used to match the candidate vaccine 2076

regimen. Due to distinctive visual characteristics or markings on presentations of licensed 2077

vaccines it may not be possible to wholly maintain double-blind conditions. In this case 2078

those site staff who prepare and/or administer trial vaccines should not otherwise be 2079

involved in trial conduct. Difficulties may also arise if the candidate vaccine is injected in a 2080

different fashion (i.e. subcutaneous, intradermal, intramuscular) to the only suitable licensed 2081

vaccine(s) that could be given to controls. In this case it may be possible to screen the 2082

administration site to prevent vaccine recipients and care-givers observing the specific 2083

method of injection. 2084

2085

o A licensed vaccine that has an effect on the infectious disease to be prevented only when due 2086

to some of the total serotypes or subtypes in the candidate vaccine. In this case the licensed 2087

vaccine provides a control group that is not vaccinated against the additional types in the 2088

candidate vaccine (i.e. unshared types). 2089

2090

If there are major objections to use of placebo injections but there is no potentially beneficial 2091

licensed vaccine that would be suitable for the target age group, the control group may be 2092

randomized to receive no vaccine. This is an undesirable situation and should be regarded as a 2093

last resort since it precludes the use of any form of blinding of trial personnel or participants 2094

(including care-givers). 2095

2096

6.2.3.2 Control groups vaccinated against the infectious disease to be prevented 2097

2098

In this case the control group receives a vaccine that is already licensed to prevent the same 2099

infectious disease as the candidate vaccine. This approach is used when it is not acceptable to 2100


employ a control group that is not vaccinated against the infectious disease to be prevented 2101

because there is at least one available licensed efficacious vaccine that is recommended for use 2102

in areas where the disease occurs. 2103

2104

On occasion, the control group receives a vaccine that may prevent the same infectious disease 2105

as the candidate vaccine but only when due to some of the total serotypes or subtypes in the 2106

candidate vaccine. Therefore the control group is vaccinated against the shared types but is not 2107

vaccinated against the unshared types. 2108

2109

If there is more than one licensed vaccine that could be used it is important that selection of the 2110

control vaccine takes into account the available evidence supporting its efficacy and, if relevant, 2111

whether it appears to have similar efficacy against all serotypes or subtypes of the pathogen 2112

involved. It is also necessary to discuss the choice of comparator with NRAs in countries where 2113

the sponsor will seek a licence for the candidate vaccine to ascertain the acceptability of an 2114

estimate of relative efficacy against a product that may be unlicensed or, at least, not the product 2115

in widespread use. This is especially important if one multi-country pivotal trial will be 2116

conducted, in which case the same vaccine should be given to the control group at all trial sites. 2117

If it is not possible to use the same control vaccine in all regions where efficacy is to be 2118

evaluated consideration should be given to conducting different efficacy trials with different 2119

vaccines used in the control groups. 2120

2121

On occasion, there may be at least one licensed vaccine available in one or more countries to 2122

prevent the same infectious disease as the candidate vaccine but there may be other countries in 2123

which the disease of interest occurs in which: 2124

o No such vaccine is yet licensed and/or 2125

o No such vaccine is included in the routine immunization schedule and/or 2126

o There are sound reasons to consider that no licensed vaccine is likely to provide useful 2127

efficacy (e.g. because the licensed vaccine does not cover or is known/expected to have 2128

poor efficacy against the serotypes or subtypes that are most prevalent in a specific 2129

region). 2130

In these situations, after careful consideration by all interested parties (i.e. sponsor, NRAs, ethics 2131


committees, local public health authorities and investigators) it may be deemed appropriate to 2132

use a control group that is not vaccinated against the disease to be prevented. 2133

2134

6.2.4 Trial designs 2135

2136

6.2.4.1 Randomization 2137

2138

The unit of randomization is most often the individual. Alternatives include the household or the 2139

cluster under trial (e.g. a school population or a local community). Randomization of groups or 2140

clusters rather than individuals may be preferred: 2141

o When a vaccination program is to be conducted in a geographical area or community 2142

o When it is logistically easier to administer the vaccine to groups than to individuals 2143

o When vaccination is anticipated to reduce transmission of the infectious agent 2144

2145

6.2.4.2 Types of trial design 2146

2147

The absolute protective efficacy of a vaccine is most commonly assessed in prospective 2148

randomized trials that compare rates of clinically apparent disease (e.g. an acute clinical illness) 2149

or established infection (e.g. chronic infection that is known to predispose to serious clinical 2150

disease) between a candidate vaccine group and a control group. 2151

2152

The simplest design involves randomization of equal numbers of subjects to each of the 2153

candidate vaccine group and the control group (i.e. 1:1). In trials that employ a control group that 2154

is not vaccinated against the disease to be prevented but there are clinical data already available 2155

to strongly support the likely efficacy of a candidate vaccine, it may be appropriate (subject to 2156

statistical considerations and an assessment of the impact on the total trial sample size) to use 2157

unbalanced randomization to reduce the chance that subjects will be randomized to the control 2158

group (e.g. 2:1 or 3:1 so that the majority of trial subjects receive the candidate vaccine). 2159

2160

Trials may plan to follow up trial subjects for the primary efficacy endpoint for a fixed period of 2161

time after the last dose of the primary series. The time at which the primary analysis is conducted 2162


is based both on the anticipated rate of occurrence of the primary efficacy endpoint in the control 2163

group and the feasibility of retaining subjects on trial for prolonged periods. Alternatively, based 2164

on anticipated rates of the primary efficacy endpoint in the control group and an expected or 2165

minimum desirable level of efficacy of the candidate vaccine, a case-driven approach may be 2166

taken. In this design the primary analysis is conducted once a pre-specified number of total cases 2167

(i.e. in a double-blind setting based on the anticipated numbers in test and control group required 2168

to demonstrate the projected vaccine effect) has been detected. 2169

2170

Alternative designs that allow for a comparison with a control group that is not vaccinated 2171

against the disease to be prevented, at least in the short-term, may include (but are not limited to) 2172

the following: 2173

2174

i) In a step-wedge trial the candidate vaccine is administered to pre-defined groups in a 2175

sequential fashion. Each pre-defined group is a unit of randomization. These may be 2176

geographical groups or groups defined by host factors (e.g. age) or other factors (e.g. attendance 2177

at a specific school or resident within a specific healthcare catchment area). Such a design may 2178

be chosen when there is good reason to anticipate that the vaccine will do more good than harm 2179

(affecting the equipoise associated with randomization to a control group that is not vaccinated 2180

against the disease to be prevented) and/or when it is impossible to deliver the intervention 2181

simultaneously to all trial participants. This design may also be used to evaluate vaccine 2182

effectiveness (see Subsection 6.3). 2183

2184

ii) In a ring vaccination trial the direct contacts of a case, and sometimes secondary contacts, 2185

may be randomized to vaccine or control or may be randomized to receive immediate 2186

vaccination or vaccination after a delay period (20). This type of pre-exposure cohort trial 2187

usually requires smaller sample sizes than prospective randomized controlled trials. The trial 2188

design assumes that there is an equal chance of vaccinees and non-vaccinees being infected and 2189

developing the infectious disease as a result of contact with an index case. 2190

2191

These types of trials may be particularly applicable when the infectious disease to be prevented 2192

is associated with a relatively high incidence of secondary cases in susceptible populations. 2193


Therefore the use of this trial design requires prior knowledge of the infectivity of the infectious 2194

agent and proportion of infections that are clinically apparent as well as the general susceptibility 2195

of the trial population. 2196

2197

The follow-up period for subjects after contact with the index case should cover the upper limit 2198

of the incubation period, taking into account the period during which the index cases were 2199

infectious and the contact period. The inclusion period for new cases and controls and their 2200

contacts should be set at a maximum of six months following the detection of the first case. 2201

Inclusion over a longer period may introduce bias in favour of vaccine efficacy, because the 2202

exposure to the infecting pathogen and thus the risk of infection will be reduced in the 2203

vaccinated groups or clusters compared with that in groups or clusters that are not vaccinated 2204

against the disease to be prevented. 2205

2206

iii) There are some situations in which the vaccine is not intended, or at least not primarily 2207

intended, to protect the vaccinees themselves against a clinically apparent infectious disease. The 2208

most common example is the vaccination of mothers during the last trimester of pregnancy, 2209

when IgG most efficiently crosses the placenta, to protect the infant during the early months of 2210

life (see Subsection 5.6.4). This strategy may or may not be followed by active immunization of 2211

infants, provided that suitable vaccines exist. If vaccine efficacy is measured in infants the unit 2212

of randomization is the mother. 2213

2214

6.2.5 Clinical endpoints 2215

2216

Preliminary efficacy trials may have an objective to identify the primary and/or secondary 2217

endpoints for confirmatory trials. Therefore the primary endpoint in preliminary efficacy trials 2218

may be different to that selected for confirmatory efficacy trials. 2219

2220

6.2.5.1 Primary endpoints 2221

2222

In most instances, the focus of vaccine efficacy trials is on the prevention of clinically apparent 2223

infections that fit the primary case definition based on clinical and laboratory criteria. The 2224


primary endpoint is also usually defined by the timeframe in which the case occurred in relation 2225

to dosing. 2226

2227

If an organism is able to cause a range of infections (e.g. from life-threatening invasive 2228

infections to common infections that are not serious if adequately treated), the primary endpoint 2229

in any one trial should be carefully selected in accordance with the proposed indication(s). 2230

2231

A candidate vaccine may contain antigens derived from one or several types (serotypes, subtypes 2232

or genotypes) of the same species. It is also possible that there may be some potential for cross-2233

protection against types not included in the vaccine (e.g. as observed with rotavirus vaccines and 2234

human papilloma virus vaccines). For these types of vaccines it is usual that the primary 2235

endpoint comprises cases due to any of the types included in the vaccine and the trial is powered 2236

for this composite endpoint. It is not usually possible to power the trial to formally assess 2237

efficacy against individual types in the vaccine or to assess cross-protection against types not in 2238

the vaccine. 2239

2240

Alternative primary endpoints may include: 2241

o Clinical manifestations of latent infection (e.g. herpes zoster) 2242

o Established chronic infections that may be asymptomatic but predispose to infection-related 2243

disease later in life (e.g. chronic hepatitis B infection; persistent infection with HPV) 2244

o Other markers that predict progression to clinically apparent disease (e.g. histological changes 2245

that are established pre-cursors of malignant neoplasia) 2246

2247

6.2.5.2 Secondary endpoints 2248

2249

As applicable to the individual candidate vaccine and the definition of the primary endpoint, 2250

important secondary endpoints may include: 2251

Cases that occur after each dose, when the vaccine schedule includes multiple doses 2252

and/or a booster 2253

Cases due to each of the individual types of the species included in the vaccine 2254


Cases due to the species (i.e. regardless of whether caused by types that are and are not 2255

included in the candidate vaccine) 2256

Cases due to non-vaccine types 2257

Cases according to host factors (e.g. age, region) 2258

Cases meeting various criteria reflecting disease severity 2259

Duration and/or severity of the illness, which may include clinical (e.g. duration of 2260

fever or rash) and laboratory measurements (e.g. duration of shedding) 2261

2262

In accordance with Subsection 5.4, one important secondary objective should be to attempt to 2263

identify a correlate of protection or, at least, a threshold value. 2264

2265

There are no vaccines indicated for the prevention or interruption of carriage, implying an effect 2266

on transmission. In addition, there are no vaccines indicated for prevention of transmission. 2267

Eradication of carriage and/or reduction in disease transmission that is not directly linked to 2268

and/or accompanied by a clinical benefit of vaccination to the individual is not usually 2269

considered to be sufficient to support licensure. Sponsors contemplating trials in which these are 2270

primary endpoints are advised to consult widely with NRAs. 2271

2272

6.2.6 Case definition 2273

2274

As part of the pre-defined primary efficacy endpoint the protocol should describe the clinical and 2275

laboratory criteria that must be met to define a case. 2276

o If a case is a clinically apparent infection it is essential that the definition includes core 2277

clinical features. It should also list acceptable sampling and laboratory processing 2278

methods to confirm the presence of the target pathogen and/or to detect infection by 2279

serological findings. 2280

o If the endpoint is the result of infection (e.g. evidence of persistence of infection or a 2281

histological change) then details of sampling (frequency and method) and grading (if 2282

applicable) should be included. 2283

2284

Adequate case definitions should also be provided for secondary endpoints. For example, if the 2285


primary endpoint is all clinically apparent infections due to the types in the vaccine the 2286

secondary analyses may focus on cases that meet specific criteria for severity, cases that require 2287

medical contact or hospitalization and cases that are due to organism types not actually included 2288

in the vaccine. 2289

2290

Whenever possible, centralized laboratories should be used and standard shipping procedures 2291

should be established for samples. If this is not feasible then information on assay performance 2292

between laboratories should be obtained and presented. The sensitivity, specificity and 2293

reproducibility of all the methods used should be included in the trial reports. If no well-2294

validated methods for establishing infection and/or progression of infection exist during the 2295

period of pre-licensure clinical development then experimental laboratory methods could be 2296

used. It would usually be expected that these experimental methods are validated before using 2297

them to analyse specimens obtained during the pivotal trials. 2298

2299

See Subsection 4.1.2 regarding the use of an adjudication committee. 2300

2301

6.2.7 Case ascertainment 2302

2303

It is critical that the same methodology for case detection is applied in all treatment groups and 2304

throughout the duration of the trial. Active case ascertainment usually requires frequent 2305

monitoring and contact with vaccinees or their care-givers. Passive case ascertainment is usually 2306

based on vaccinees or care-givers presenting to or otherwise contacting a local healthcare facility 2307

due to the onset of specific symptoms. In this case it is common that contact is triggered by one 2308

or more of a list of signs or symptoms given to trial subjects or their care-givers at the time of 2309

randomization and they may be instructed to contact a specific healthcare facility. Alternatively 2310

or in parallel, cases may be detected based on monitoring all local clinics and hospitals for cases. 2311

2312

For efficacy endpoints based on clinically apparent disease, the possible range of clinical 2313

presentations will determine the mode of case ascertainment. For example, this may be hospital-2314

based for cases of life-threatening infections or community based for less severe infections. If 2315

community based, case detection may depend on family practitioners and on first suspicion of 2316


infection by vaccinated subjects themselves or their parents/guardians. In each case, it is 2317

critically important that the individuals who are most likely to initiate detection of a possible 2318

case should have clear instructions. These may need to cover issues such as criteria for 2319

stimulating contact with designated healthcare professionals, telephone contacts, initial 2320

investigations and further investigations once a case is confirmed. 2321

2322

For efficacy endpoints other than clinically apparent disease, it becomes critical that subjects are 2323

monitored at regular intervals to detect clinically non-apparent infections or changes in other 2324

selected markers (e.g. the appearance of histological changes). The frequency of visits, and 2325

acceptable windows around the visits, should be laid down in the trial protocol and must be 2326

carefully justified. 2327

2328

The appropriate period of case ascertainment during a trial requires special attention and will be 2329

determined mainly by the characteristics of the disease to be prevented and the claim for 2330

protection that is sought at the time of initial authorization. For infectious diseases that have 2331

marked seasonality, at least in some geographic locations, it is usual to plan for a primary 2332

analysis at least when all vaccinees have been followed through one complete season. In these 2333

settings it is usual to conduct an enrolment campaign over a very short period just before the 2334

expected season onset. However, it may be necessary to repeat the exercise before the next 2335

season to meet the pre-defined sample size, in which case the opportunity should be taken to 2336

collect all cases that occur in the second season for the initial vaccination campaign cohort. 2337

2338

6.2.8 Duration of follow-up 2339

2340

At the time of conducting the primary analysis for the purposes of obtaining initial licensure, the 2341

duration of follow-up in vaccine efficacy trials may be relatively short (e.g. 6-12 months) and 2342

insufficient to detect waning protection, if this exists. Therefore, case ascertainment should 2343

continue in the vaccine efficacy trial populations and/or waning protection should be assessed 2344

during post-licensure effectiveness trials. These data may serve to indicate the need for and 2345

optimal timing of booster doses and to estimate efficacy after booster doses. 2346

2347


6.2.9 Analysis of efficacy 2348

2349

6.2.9.1 Sample size calculation 2350

2351

The trial sample size should be calculated based on: 2352

i. The selected primary efficacy endpoint, including the possibility that the primary 2353

endpoint may be a composite of cases due to any of the organism types included in 2354

the candidate vaccine; 2355

ii. The primary analysis population (see below) and 2356

iii. According to the primary hypothesis (i.e. superiority or non-inferiority and the pre-2357

defined criteria). 2358

2359

If the primary analysis population represents a subset of the total randomized population the 2360

sample size calculation should include an adequate estimation of numbers likely to be excluded 2361

from the primary analysis for various reasons. In addition, if considered necessary, a blinded 2362

review of total numbers enrolled who are eligible for the primary analysis population may be 2363

conducted after a pre-defined number has been randomized so that the trial sample size can be 2364

adjusted accordingly. 2365

2366

6.2.9.2 Analysis populations 2367

2368

Clinical efficacy is usually assessed in the total randomized trial population (i.e. those who are 2369

assigned to receive vaccine and/or control) and in pre-defined subsets of the randomized 2370

population. 2371

2372

In maternal immunization trials of clinical efficacy it may be appropriate that trials are powered 2373

to assess vaccine efficacy only in the offspring. If a secondary or exploratory analysis is 2374

conducted in mothers the case definition will likely need to be different. 2375

2376

The pre-defined trial populations should include as a minimum: 2377

o All randomized subjects (i.e. the full analysis set) 2378


o All vaccinated subjects regardless of the numbers of assigned doses actually received and 2379

whether or not they were administered within the pre-defined windows 2380

o Subsets of all vaccinated subjects separated according to any evidence of prior exposure 2381

to the infectious disease under trial (e.g. baseline seropositivity vs. seronegativity) 2382

o The per protocol population should be confined to subjects who have generally complied 2383

with the protocol and have received all assigned doses within pre-defined windows. In 2384

addition, this population should be confined to those with no evidence of prior 2385

exposure to the infectious agent (or specific serotypes or subtypes) at baseline. 2386

Depending on the target pathogen this subset may also be defined based on prior 2387

vaccination history. 2388

2389

Other populations may be appropriate for some pre-defined secondary or exploratory analyses. 2390

For example: 2391

o Those who completed specific numbers of assigned doses or received all doses within 2392

pre-defined windows around the scheduled trial visits, i.e. analyses of efficacy according 2393

to adherence to the vaccination regimen 2394

o Subgroups defined by demographic factors known or postulated to impact on vaccine 2395

efficacy 2396

2397

6.2.9.3 Primary analysis 2398

2399

It is common in vaccine efficacy trials that the pre-defined primary analysis is based on 2400

estimating efficacy in the per protocol population and on rates of true vaccine failures, i.e. the 2401

calculation of efficacy takes into account only those cases with onset after a minimum time had 2402

elapsed after completion of the assigned doses. For example, depending on knowledge of the 2403

kinetics of the immune response, true vaccine failures may be limited to cases with onset more 2404

than a specified number of days or weeks after the final dose of the primary series. In addition, 2405

for a vaccine that contains antigens from only certain serotypes or subtypes, the primary analysis 2406

may be based on cases due to vaccine types only. 2407

2408

In trials that compare a candidate vaccine with a group that is not vaccinated against the disease 2409


to be prevented the aim is to demonstrate that the lower bound of the 95% confidence interval 2410

around the estimate of vaccine efficacy is above a pre-defined percentage (which will always be 2411

above zero). The pre-defined percentage should be selected based on the sponsor’s expectation 2412

of the point estimate of vaccine efficacy and taking into account what might be viewed as the 2413

minimum level of efficacy that could be considered clinically important. The sample size 2414

calculation is based on this objective. 2415

2416

In trials that compare a candidate vaccine with an active control the aim is to demonstrate non-2417

inferiority of the candidate vs. the control vaccine, with calculation of the 95% confidence 2418

intervals around the difference in rates of breakthrough infections. This requires a pre-defined 2419

non-inferiority margin, which should be justified in accordance with prior estimates of vaccine 2420

efficacy for the disease to be prevented, and level of alpha on which the sample size calculation 2421

depends. If the sponsor also intends to assess superiority of the candidate vaccine over the active 2422

control the statistical analysis plan should pre-define a hierarchical assessment so that superiority 2423

is assessed only after establishing that the non-inferiority has been demonstrated. 2424

2425

6.2.9.4 Other analyses 2426

2427

The full range of secondary and exploratory analyses will depend on the pre-defined endpoints. 2428

Some of these analyses may be conducted in specific predefined trial populations. For example, 2429

important sensitivity analyses to support the primary analysis include those based on all proven 2430

cases whenever they occurred after randomization and in each analysis population. If the 2431

schedule includes more than one dose then analyses should be conducted that count cases from 2432

the time of each dose for all subjects who were dosed up to that point. 2433

2434

If the primary analysis was confined to cases due to organism types included in the vaccine then 2435

additional analyses should evaluate efficacy based on all cases regardless of the serotype or 2436

subtype responsible. If there are sufficient numbers of cases, these analyses may provide some 2437

indication of any cross-protection provided by the antigens in the vaccine. 2438

2439

Depending on the case definition, other analyses may be based on cases that met some but not all 2440


of the case definition criteria, cases that were severe and cases that required a medical 2441

consultation or hospitalization. 2442

2443

6.2.9.5 Other issues 2444

2445

Vaccines that contain antigens derived from several serotypes, subtypes or genotypes 2446

2447

As discussed in Section 4.3.5, it is not usually possible to power the trial to formally assess 2448

efficacy against individual types in the vaccine. Secondary or, at least, exploratory analyses 2449

should be planned to describe efficacy against the various types represented in the vaccine and, if 2450

there is an expectation of cross-protection, against types not included. If the data suggest 2451

unusually low efficacy against any type in the vaccine it may be necessary to explore this matter 2452

in further trials. 2453

2454

Magnitude of vaccine efficacy 2455

2456

The point estimate of vaccine efficacy and 95% confidence intervals that are obtained may 2457

indicate that a relatively modest proportion of cases can be prevented. This fact alone does not 2458

preclude licensure provided that the sponsor can substantiate that the vaccine efficacy observed 2459

represents an important clinical benefit. For example, if the vaccine prevents life-threatening 2460

infections for which there is no very effective specific therapy and for which no vaccine or no 2461

more effective vaccine is available. 2462

2463

Extrapolation of vaccine efficacy 2464

2465

Vaccine efficacy can only be estimated in geographical areas where there is sufficient disease to 2466

support trial feasibility. In most instances it is not necessary for any one NRA to request 2467

provision of efficacy data from within its own jurisdiction nor is it feasible to conduct a study 2468

that provides robust results within a single country. Any such requests should only be made 2469

when there are scientifically sound reasons to think that vaccine efficacy could be substantially 2470

lower compared to that observed in the areas where Phase 3 trials were conducted. In addition, 2471


such requests should not be made if there is a good scientific justification to use immunobridging 2472

to support extrapolations of efficacy between populations (see Section 5 on bridging efficacy). 2473

2474

6.3 Approaches to determination of effectiveness 2475

2476

Vaccine effectiveness reflects direct (vaccine induced) and indirect (population related) 2477

protection during routine use. Thus, the assessment of vaccine effectiveness can provide useful 2478

information in addition to any pre-authorization estimates of protective efficacy. Even if it was 2479

not feasible to estimate the protective efficacy of a vaccine pre-authorization it may be possible 2480

and highly desirable to assess vaccine effectiveness during the post-authorization period. The 2481

information gained from assessments of vaccine effectiveness may be particularly important to 2482

further knowledge on the most appropriate mode of use of a vaccine (e.g. need for booster doses 2483

in at least some segments of the population to maintain adequate protection over time). 2484

2485

Vaccine effectiveness may be estimated: 2486

i) In observational cohort trials that describe the occurrence of the disease to be prevented in 2487

the target population over time. However, there is no randomization step and there is the 2488

potential for considerable biases to be introduced. One such approach is the screening 2489

method. 2490

ii) During phased (e.g. in sequential age or risk groups) introduction of the vaccine into the 2491

target population in which the groups might form the units of randomization (i.e. using a 2492

stepped wedge design). 2493

2494

iii) Using other designs, of which a wide range has been used in different circumstances. For 2495

example, using a case test-negative trial design. In this modification of a case control trial 2496

subjects with symptoms suggesting the infectious disease under trial and seeking medical 2497

care are tested for the infectious agent of interest. The cases are those who are positive and 2498

controls are those who are negative for the pathogen of interest. If vaccinated cases are less 2499

severely ill and seek care less frequently than cases that occur in individuals not vaccinated 2500

against the disease to be prevented, then an appropriate adjustment for illness severity is 2501

required to avoid bias in effectiveness estimates (21). 2502


2503

Vaccine effectiveness is affected by a number of factors, including: 2504

o Vaccination coverage of the population 2505

o Pre-existing immune status of the population 2506

o Differences in types included in a vaccine compared to predominant circulating types 2507

o Changes in circulating predominant types over time 2508

o Transmissibility of the pathogen and any effect that introduction of routine vaccination 2509

may have had on transmission rates 2510

2511

It may not be possible or appropriate for sponsors to conduct trials to estimate vaccine 2512

effectiveness themselves since regional or national networks may be necessary to ensure that 2513

cases are reliably detected. For some types of disease the use of data collected by means of 2514

national or international registries may be appropriate. In addition, in some jurisdictions the 2515

estimation of vaccine effectiveness is not considered to fall within the remit of the license holder. 2516

2517

Whatever the local requirements and arrangements, sponsors should discuss the arrangements for 2518

ongoing disease surveillance and the potential for estimating effectiveness with public health 2519

authorities in countries where the vaccine is to be used and where appropriate surveillance 2520

systems are in place. The plans for estimation of effectiveness should also be agreed with NRAs 2521

at the time of licensure and the requirements for reporting of effectiveness data to the NRA 2522

either via the sponsor or directly from a public health authority should be clarified. 2523

2524

It may be that reliable estimates of effectiveness can only be obtained in certain countries in 2525

which vaccination campaigns are initiated and where there is already a suitable infrastructure in 2526

place to identify cases. Therefore, it would likely be inappropriate to extrapolate any estimates of 2527

effectiveness that are obtained to other modes of use (such as introducing the same vaccine to 2528

different or only to highly selected sectors of the population). 2529

2530

7. Safety 2531

2532



Evaluating safety in clinical trials 2534

- Safety as a primary or secondary endpoint 2535

- Recording and categorisation of adverse events within trials 2536

- Size of the pre-licensure safety database 2537

Post-licensure safety surveillance 2538

- Spontaneous reporting 2539

- Roles of the license holders and NRAs 2540

2541


2543

Safety should be assessed in all clinical trials that are conducted pre- or post-licensure. The 2544

assessment of safety may be the only primary objective, a co-primary objective or a secondary 2545

objective in a clinical trial. Since the methods for collection, analysis and interpretation of safety 2546

data during clinical trials contrast with those applicable to post-licensure routine safety 2547

surveillance they are considered separately. 2548

2549

In principle, many of the approaches to documenting and reporting safety data during clinical 2550

trials and the conduct of pharmacovigilance activities for vaccines are similar to those for all 2551

medicinal products. The sections that follow should be read in conjunction with the extensive 2552

guidance that is available from many publications and on the websites of WHO, CIOMS, the 2553

ICH and individual regulatory bodies. The focus of the sections is on some methods and 2554

practises that are different for vaccines compared to other medicinal products and on some issues 2555

that may need to be addressed due to the vaccine composition. 2556

2557

7.2 Assessment of safety in clinical trials 2558

2559

As described in Subsection 4.1.2 the use of a DSMB should be considered before commencing 2560

clinical trials. If the DSMB’s role includes recommending early termination of a trial there 2561

should be appropriate stopping rules in place. 2562

2563

7.2.1 Safety as a primary or secondary endpoint 2564


2565

7.2.1.1 Safety as a primary endpoint 2566

2567

In the early clinical trials with a new candidate vaccine the assessment of safety may be the only 2568

primary objective or a co-primary objective. It is very unusual that the assessment of safety is a 2569

primary objective in pre-licensure trials conducted later in the development program. Where this 2570

has occurred the focus has been on a specific safety issue (e.g. intussusception in pre-licensure 2571

trials with rotavirus vaccines that were developed after the first vaccine had indicated a potential 2572

association with vaccination). The assessment of one or more safety aspects is the primary 2573

objective in post-licensure safety trials, which involve detailed monitoring during routine 2574

immunization programs. 2575

2576

When the assessment of safety is the primary objective of a clinical trial it is usual that the 2577

primary analysis is based on a specific safety endpoint (e.g. rates of a certain adverse event [AE], 2578

rates of AEs within a specific system organ class [SOC] or rates of AEs that may be part of a 2579

clinical syndrome of interest). These trials should be powered to address the pre-specified 2580

hypothesis. The exception is in trials that are exploratory in nature, such as initial trials with new 2581

candidate vaccines intended to provide a preliminary assessment of the safety of ascending doses 2582

or sequential doses. 2583

2584

7.2.1.2 Safety as a secondary endpoint 2585

2586

In vaccine efficacy trials and in immunogenicity trials the assessment of safety is usually a 2587

secondary objective. These trials are not powered a priori to support formal statistical 2588

conclusions from analyses of rates of all or specific AEs between trial groups but simple 2589

statistical comparisons are commonly used as an initial screening for any differences in rates 2590

between groups of subjects. If such analyses are conducted they should be pre-specified in the 2591

protocol and in the statistical analysis plan. If there are any findings indicating statistically 2592

significant differences in rates of AEs (overall, by SOC or by PT) they need to be interpreted 2593

with caution due to the fact that the trial was not primarily designed to address pre-specified 2594

hypotheses regarding safety endpoints. Nevertheless, the findings may indicate that it is 2595


appropriate to design and power further pre- or post-licensure clinical trials to further investigate 2596

and quantify the potential risks. 2597

2598

7.2.2 Recording and reporting adverse events 2599

2600

7.2.2.1 Methods 2601

2602

Adverse events and serious adverse events (SAEs) should be reported and recorded by 2603

investigators and sponsors according to detailed procedures described in the trial protocol and in 2604

accordance with requirements for expediting reporting to NRAs. 2605

2606

In safety and immunogenicity trials it is usually expected that all AEs, whether solicited or 2607

unsolicited, are collected for defined periods after each dose from all randomized subjects or all 2608

randomized subjects who received at least one dose of assigned treatment (see Subsections 2609

7.2.2.2 and 7.2.2.3). In vaccine efficacy trials involving large numbers of subjects, taking into 2610

account the safety profile observed in the previous trials and the numbers from which detailed 2611

safety data have already been obtained, it may be acceptable that all AEs are collected from a 2612

randomized subset. In this case all SAEs and any pre-specified adverse events of special interest 2613

(AESIs) should be collected from all randomized subjects. It may also be acceptable that only 2614

SAEs and AESIs are collected during long-term safety follow-up. 2615

2616

7.2.2.2 Solicited signs and symptoms 2617

2618

After each dose of a vaccine or placebo, local and systemic solicited signs and symptoms should 2619

be documented for a pre-defined post-dose period by vaccinees or their care-givers by 2620

completing a daily diary record. These diaries should be filled in each day and users should 2621

receive instructions in their completion before vaccination commences. The duration of 2622

collection of data in diaries should be at least 5-7 days after each dose but longer periods (e.g. 2623

10-14 days) may be appropriate for vaccines that contain live micro-organisms, depending on 2624

whether or not they are replication-competent. 2625

2626


For injectable vaccines the local signs and symptoms to be documented are usually pain, redness 2627

and swelling in all age groups. When two or more vaccines are given by injection at the same 2628

time, the diary card should ensure that separate data are recorded for each injection site (for 2629

example, these are usually into different limbs and therefore the diary card should contain 2630

separate records by right and left arm and/or leg). For vaccines given by other routes, alternative 2631

local signs and symptoms may be identified as representing local AEs (e.g. sneezing after 2632

intranasal dosing). The systemic signs and symptoms are determined by the age range in the trial 2633

(e.g. those appropriate for infants will not be wholly applicable to toddlers and older subjects) 2634

and the route of administration (e.g. nausea and vomiting could be solicited symptoms for 2635

vaccines given orally). 2636

2637

For subjective symptoms (e.g. pain, fatigue, myalgia) a simple scoring system should be 2638

included in the diaries to allow for a grading of severity. For objective signs, the quality of the 2639

information collected can be improved by methods such as issuing digital thermometers to each 2640

vaccinee or care-giver for application at a specific site (e.g. oral or axillary in infants, with 2641

recordings made at specific time of the each day) and using transparent plastic measuring 2642

devices to record the extent of redness and swelling. 2643

2644

Any self-administered treatments used to address signs or symptoms (such as antipyretic and 2645

analgesic medicines) and whether there was any contact with, or treatment administered by, a 2646

healthcare professional should be captured. If a supply of a specific anti-pyretic or analgesic was 2647

given out at the time of each dose for use as needed, or as instructed in accordance with the 2648

protocol, the post-dose usage recorded in the diary should be checked against returned supplies. 2649

If prior safety data suggest that pre-vaccination antipyretic use is appropriate, this can be 2650

administered and recorded by trial staff at the vaccination visit and the diary cards should collect 2651

any post-vaccination doses administered. 2652

2653

At each trial visit, whether it involves face-to-face or telephone contact between the vaccinee 2654

and/or care-giver and trial staff, the diary cards should be checked for level of completion and 2655

further instructions given as needed to improve data recording after the next dose is given. At 2656

face-to-face visits the prior vaccination site(s) should be inspected for any remaining signs such 2657


as induration. Also, vaccinees or care-givers should be asked about the maximum extent of signs 2658

(e.g. to determine whether whole limb swelling occurred). Any unresolved local or systemic 2659

signs and symptoms should be recorded and action taken as appropriate. 2660

2661

7.2.2.3 Unsolicited AEs 2662

2663

In addition to signs and symptoms that are pre-specified for collection of data, vaccinees and/or 2664

their care-givers should be questioned at each trial visit for the occurrence of any AEs since the 2665

last visit. For each AE the timing of onset in relation to vaccination, whether a healthcare 2666

professional was consulted, whether hospitalisation occurred and any treatment that was given 2667

(prescribed or non-prescribed) should be captured. Sponsors may also wish to record any days 2668

off school or off work for vaccinees and days off work for their care-givers. 2669

2670

A checklist of symptoms that could possibly reflect the onset of a pre-specified AESI may be 2671

useful to identify potential cases of various syndromes (such as auto-immune diseases) at an 2672

early stage and to ensure that there is careful follow-up. In addition, questions should be posed to 2673

elicit whether certain AEs have occurred that could be anticipated in the age group studied. For 2674

example, to determine whether persistent inconsolable crying or hypotonic hypo-responsive 2675

episodes occurred in infants. Where well-established and widely-applied definitions of these and 2676

other AEs are available, the reports received should be classified using these criteria. 2677

2678

Although solicited signs and symptoms are AEs, it is usual that clinical trial reports tabulate 2679

safety data separately for these and for unsolicited AEs. The classification of AEs should use a 2680

standardised scheme, such as MedDRA, to categorise AEs by SOC and PT. If the classification 2681

scheme is updated during conduct of the trial the clinical trial report should indicate how the 2682

changes impact on the tabulations. 2683

2684

7.2.2.4 Other investigations 2685

2686

The collection of data on routine laboratory tests (haematology, chemistry and urinalysis) is not 2687

commonly perceived to be necessary in clinical trials with vaccines. If the sponsor or NRA 2688


considers that there is a good rationale for obtaining these data at certain time points the results 2689

should be generated in appropriately certified laboratories and reported using well-established 2690

grading scales for abnormalities. 2691

2692

For vaccines that contain live organisms (including attenuated wild-types, organisms that have 2693

been genetically engineered to render them non-virulent and/or non-replicative and live viral 2694

vector vaccines) additional investigations related to safety should usually include the detection of 2695

viraemia and assessments of shedding (quantity and duration). Organisms recovered from 2696

vaccinees may also be subjected to genetic analyses to determine any instances of recombination 2697

with wild types and reversion to virulence and/or replication competency. 2698

2699

In the case of vaccines administered to pregnant women measures of growth and development in 2700

their infants may be important safety parameters. 2701

2702

7.2.3 Categorization of adverse events 2703

2704

7.2.3.1 Causality 2705

Section 8.5 of the WHO Global Manual on Surveillance of Adverse Events Following 2706

Immunization (22) recommends that in clinical trials the investigator should make a judgement 2707

of relatedness to vacination for all solicited signs and symptoms and unsolicited AEs. The 2708

investigator’s assessment may also be commented on by the sponsor. The assessment of 2709

relatedness to vaccination should take into account factors such as: 2710

a) Plausibility of relatedness, taking into account the vaccine construct. For example, live 2711

attenuated vaccines may be associated with modified manifestations of natural infection 2712

(e.g. rashes). 2713

b) Timing in relation to dosing. Whilst most vaccine-related AEs occur within 1-2 weeks after 2714

a dose there may reasons to suspect that illnesses with onset many months after the last dose 2715

could be related to prior vaccination. For example, for some powerful adjuvants there is a 2716

hypothetical concern that rates of auto-immune diseases may increases in genetically-2717

prediposed sub-populations. 2718


c) Concurrent illnesses common in the trial age group or documented in the case report form 2719

and the anticipated background rates, if known. This is a particular issue for vaccines 2720

administered to infants and young children in whom intercurrent illnesses are relatively 2721

common. 2722

d) The frequency with which any one AE occurred in groups that received the candidate 2723

vaccine compared to groups that received another vaccine or placebo. 2724

e) Any correlation between rates of any one AE and dose of antigenic components. 2725

f) Changes in rates of any one AE with sequential doses. 2726

g) The results of medical investigations (e.g. diagnostic tests for concurrent illnesses) and of 2727

autopsies (e.g. in cases of sudden infant death). 2728

2729

7.2.3.2 Severity 2730

2731

Sufficient data should be collected for each solicited sign and symptom and unsolicited AE to 2732

make an assessment of severity. Wherever possible widely used grading scales should be used 2733

and/or the same scales should be applied throughout the clinical development program. 2734

2735

7.2.3.3 Other categorization 2736

2737

The classification of AEs as serious and the categorisation of frequencies should follow 2738

internationally-accepted conventions, as described in Section 3.1.2 of the WHO Global Manual 2739

on Surveillance of Adverse Events Following Immunization (22). Frequencies of solicited signs 2740

and symptoms by subject and of AEs in each treatment group should be calculated based on the 2741

denominator of all vaccinated subjects in that group. Frequencies of solicited signs and 2742

symptoms after each dose should use the number that received each dose. 2743

2744

7.2.4 AE reporting rates within and between trials 2745

2746

During any one clinical development program the reporting rates for all and/or for specific types 2747

of AEs, whether solicited or unsolicited, in clinical trials may demonstrate: 2748

2749


i) Differences between candidate vaccines and control groups within a clinical trial. For 2750

example, differences in AE rates may be anticipated between a candidate vaccine and a 2751

placebo group or a group that receives a licensed vaccine that does not have a similar 2752

composition to the candidate vaccine. Any marked differences between a candidate vaccine 2753

and a licensed vaccine that has the same or very similar composition are generally not 2754

anticipated and may require further investigation. 2755

2756

ii) Differences between clinical trials that may be observed in one or both of the candidate 2757

vaccine and control groups for total or specific AE reporting rates. Whenever this occurs it 2758

is important to consider the possible explanations, taking into account whether or not the 2759

same effect on the pattern of reporting rates is observed in groups that receive candidate 2760

vaccines and licensed vaccines and whether the study was double-blind or open-label. These 2761

differences between trials may reflect real and anticipated differences in vaccine 2762

reactogenicity between trial populations (e.g. age-related differences for specific AEs, such 2763

as higher fever rates in trials conducted in infants and toddlers compared to those in older 2764

children and adults). In contrast, marked differences in reporting rates between trials 2765

conducted in similar age ranges but in different geographical locations would not usually be 2766

anticipated. When there is no clear explanation for the differences observed, consideration 2767

should be given to the possibility that there has been incomplete reporting of AEs and 2768

further investigation is merited. 2769

2770

7.3 Size of the pre-licensure safety database 2771

2772

A total database of 3000 subjects across all trials and populations provides a 95% chance of 2773

observing one instance of an AE that occurs on average in 1 in 1000 subjects. This number may 2774

be regarded as a generally applicable target for the minimum total pre-licensure safety database 2775

for a new candidate vaccine that contains one or more antigenic components not previously used 2776

in human vaccines. Nevertheless, this figure should not be applied to application dossiers for any 2777

type of new candidate vaccine without further considerations, which include the following: 2778

a. Fewer than 3000 subjects may be acceptable if the new candidate vaccine consists only of 2779

antigenic components already licensed in other vaccines for which there is considerable 2780


experience in routine use. 2781

b. The total number exposed in clinical trials may cover many age sub-groups or a single age 2782

group may predominate. It may be acceptable that the majority of subjects included in the 2783

safety database come from a specific age range unless the available data point to some 2784

specific safety concerns that require further investigation in other age groups before 2785

licensure. 2786

c. For specific types of vaccines (e.g. innovative constructs) or specific modes of use (e.g. in a 2787

population considered to be vulnerable or otherwise at high risk that could predispose them 2788

to certain adverse events) individual NRAs may require that considerably more than 3000 2789

subjects are exposed prior to initial licensure. 2790

d. Additional considerations may apply to vaccines that contain antigenic components not 2791

previously used in human vaccines but for which efficacy trials are not possible. A large 2792

pre-licensure safety database is highly desirable for a vaccine with potential to be 2793

administered to very large numbers in an emergency situation (e.g. influenza pandemic 2794

vaccines, vaccines against certain viral haemorrhagic fevers or smallpox vaccines). 2795

Nevertheless, the safety profile documented in the initial safety and immunogenicity trials 2796

may lead to some reluctance to unnecessarily expose large numbers of subjects in the 2797

absence of an immediate threat and/or to expose large numbers in particular population 2798

subsets. Therefore NRAs may consider licensing these types of vaccines based on a 2799

relatively small safety database provided that very detailed plans are in place at the time of 2800

licensure for monitoring of safety should it be necessary to give the vaccine to large 2801

numbers of individuals at some future time. 2802

2803

7.4 Post-licensure safety surveillance 2804

2805

The requirements of individual NRAs for reporting of safety data collected from post-licensure 2806

safety surveillance activities should be consulted. NRAs should provide publicly-available 2807

guidance regarding their requirements for the content and timing of periodic reports of safety 2808

data and for any expedited reporting considered necessary. License holders should demonstrate 2809

that they have adequate capability and appropriate staff to collect, interpret and act upon the 2810

safety data received. 2811


2812

It has become routine that at the time of initial licensure there are detailed proposals in place for 2813

post-licensure safety surveillance activities, often in the form of risk management plans. These 2814

documents and proposals are then routinely updated at intervals in line with additional data that 2815

become available. They usually outline the safety specification for the vaccine based on all 2816

available safety data at the time of submitting each version of the plan along with details of 2817

routine and proposed additional pharmacovigilance and risk minimisation activities. 2818

2819

When planning pharmacovigilance activities for a vaccine, it is important to take into account 2820

that in addition to routine pharmacovigilance (i.e. passive surveillance), important information 2821

may come from: 2822

i) Data from enhanced safety surveillance (active surveillance) put in place by public health 2823

bodies when a vaccine is introduced into a national routine immunization program or when 2824

the use of a vaccine within a program changes significantly (e.g. an entirely different age 2825

group is vaccinated for the first time). 2826

ii) Large databases that link information in patient records on vaccination history with 2827

occurrence of specific types of illness. These can be interrogated to explore links between 2828

specific vaccines and safety issues in the short and longer-term. 2829

iii) Various types of registries intended to capture details of use in specific populations. For 2830

example, there are registries that collect information on exposure of pregnant women to 2831

various types of vaccines and the outcome of the pregnancy (including rates of spontaneous 2832

abortion, premature delivery and congenital malformations in the infants). There are also 2833

registries that capture specific types of disease that could be of relevance to specific types of 2834

vaccines. 2835

2836

The limitations of each of these approaches are well known, which underlines the need to 2837

consider all sources along with additional data that may come from post-licensure trials. 2838

2839

As with other medicinal products the same vaccine may be marketed by different license holders 2840

in various countries and regions so that systems need to be in place at the time of licensure to 2841

facilitate rapid sharing of safety information between companies, between companies and NRAs 2842


and between NRAs. An additional consideration for vaccines is that when a safety signal is 2843

identified for any one vaccine it may or may not be possible to ascribe the AEFIs observed to 2844

any one antigenic component of the vaccine or to an adjuvant. Furthermore, if there was 2845

concomitant administration of vaccines in some or all cases generating the signal it may not be 2846

possible to ascribe the AEFI to only one of the products co-administered. The same or very 2847

similar antigenic component(s) or adjuvant in the vaccine(s) from which the signal arose may be 2848

in several other licensed products marketed worldwide. Ultimately several different companies 2849

and NRAs without established data sharing agreements may need to be involved. As a result, the 2850

actions taken, if any, and the speed at which action has been taken, are sometimes very variable 2851

between countries. These issues underscore the need for efficient use of electronic databases to 2852

facilitate rapid data sharing. 2853

2854

Authors and Acknowledgements 2855

2856

The first draft of this document was prepared by the WHO Drafting Group comprising Dr M. 2857

Powell, Medicines and Healthcare Products Regulatory Agency, London, United Kingdom; Dr 2858

R. Sheets, Consultant, Silver Spring (MD), USA; Dr John McEwen, Medical Adviser, 2859

Therapeutic Goods Administration, ACT, Canberra, Australia, Dr I. Knezevic, Department 2860

of Essential Medicines and Health Products, World Health Organization, Geneva, Switzerland, 2861

and Dr Vasee Moorthy, Initiative for Vaccine Research, World Health Organization, Geneva, 2862

Switzerland, taking into considerations the discussion and consensus reached at the WHO 2863

consultation on Clinical Evaluation of Vaccines held on 17-18 July 2014, at WHO, Geneva, 2864

Switzerland, attended by the following participants: 2865

2866

Paula Annunziato, Executive Director, Clinical Research, Merck & Co., New Jersey, United 2867

States of America, Niranjan Bhat, Senior Clinical Officer, Vaccine Access and Delivery, 2868

Program for Appropriate Technology in Health, Seattle, United States of America, Arani 2869

Chatterjee, Senior vice President, Clinical R&D, Biological E Ltd, Hyderabad, India, Keith 2870

Chirgwin, Deputy Director, Program Strategies, Bill & Melinda Gates Foundation, Seattle, 2871

United States of America, Gina Coleman, Chief, Clinical Evaluation Division, Health Canada, 2872

Ottawa, Canada, Do Tuan Dat, Director, The Company for Vaccines and Biological Production 2873


No. 1 (VABIOTECH), Ha Noi, Viet Nam, Patricia E. Fast, International AIDS Vaccine Initiative, 2874

New York, United States of America, Ginamarie Foglia, Director, Clinical Development, Sanofi 2875

Pasteur, Swiftwater, United States of America, Uli Fruth, Initiative for Vaccine 2876

Research, World Health Organization, Geneva, Switzerland, Marion Gruber, Director, Office of 2877

Vaccines Research and Review, Center for Biologics Evaluation and Research, U.S. Food and 2878

Drug Administration, Rockville, United States of America, Penny M. Heaton, Director, 2879

Vaccine Development, Bill & Melinda Gates Foundation, Seattle, United States of America, 2880

David Kaslow, Vice President, Product Development, PATH, Program for Appropriate 2881

Technology in Health, Washington DC, United States of America , Ivana Knezevic, 2882

Technologies Standards and Norms, World Health Organization, Geneva, Switzerland, Olivier 2883

Lapujade, Prequalification Team, World Health Organization, Geneva, Switzerland, Yun Hee 2884

Lee, Scientific Officer/Reviewer, Biologics Division, Biopharmaceuticals & Herbal Medicine 2885

Evaluation, Ministry of Food & Drug Safety, Chungcheongbuk-do, Republic of Korea, 2886

David J.M. Lewis, Professor of Clinical Vaccine Immunology, Clinical Research Centre, 2887

Institute of Biosciences and Medicine, FHMS, University of Surrey, Guildford, United 2888

Kingdom, Annette Lommel, Clinical Reviewer, Paul Ehrlich Institute, Langen, Germany, 2889

John McEwen, Medical Adviser, Therapeutic Goods Administration, ACT, Canberra, 2890

Australia, Vaseeharan Sathiyamoorthy, Initiative for Vaccine Research, World Health 2891

Organization, Geneva, Switzerland, Pieter Neels, Vaccine-Advice BVBA, Zoersel, Belgium, 2892

Marijke Nijs, Director, Clinical Regulatory Excellence, GlaxoSmithKline Biologicals, Wavre, 2893

Belgium, Sérgio Andrade Nishioka, Coordinator, Clinical Research, Department of Science 2894

and Technology, Ministry of Health, Brasilia, Brazil, Audino Podda, Head, Clinical 2895

Development and Regularity Affairs, Novartis Vaccines Institute for Global Health (NVGH), 2896

Siena, Italy, Mair Powell, Medicines and Healthcare products Regulatory Agency, London, 2897

United Kingdom, Ajmeer Ramkishan, Deputy Drugs Controller, Central Drugs Standard Control 2898

Organization, New Delhi, India, Rebecca Sheets, Consultant, Grimalkin Partners, Silver Spring, 2899

United States of America , Jinho Shin, Expanded Programme on Immunization, World 2900

Health Organization, Western Pacific Regional Office, Manila, the Philippines, Peter Smith, 2901

MRC Tropical Epidemiology Group, London School of Hygiene and Tropical Medicine, 2902

London, United Kingdom, James Southern, Advisor to Medicines Control Council in South 2903

Africa, Medicines Control Council, Cape Town South Africa, Yuansheng Sun, Clinical and 2904


Nonclinical assessor, Paul-Ehrlich-Institut, Langen, Germany, Kirsten Vannice, Initiative for 2905

Vaccine Research, World Health Organization, Geneva, Switzerland, David Wood, Technologies 2906

Standards and Norms, World Health Organization, Geneva, Switzerland, Zhimin Yang, Vice-2907

chief of Office, Office of Evaluation III, CDE, Beijing, People’s Republic of China. 2908

2909

The draft document was posted on the WHO Biologicals web site for the first round of public 2910

consultation from 30 October to 30 November 2015. 2911

2912

The second draft was prepared by the WHO Drafting Group, taking into account comments 2913

received from following reviewers: 2914

2915

Dr Bernard Fritzell, BFL conseils, France; Dr Grace Chen, National Institutes of Health, USA; 2916

Dr Gina Coleman, Health Canada, Ottawa, Canada; Zuzana Kusynová consolidated comments of 2917

International Pharmaceutical federation (FIP), The Netherlands; Marijke Nijs, GlaxoSmithKline 2918

Vaccines, Belgium; Novilia Sjafri and BioFarma’s Clinical Team, Indonesia; Dr Yuansheng Sun, 2919

Clinical and Nonclinical assessor, Paul-Ehrlich-Institut, Langen, Germany; Ingrid Uhnoo, 2920

Uppsala universitet, Sweden; Dr Teruhide Yamaguchi, Pharmaceutical and Medical Devices 2921

Agency, Japan; Dr Kathryn Zoon, National Institutes of Health, USA. 2922

2923

The draft document is posted on the WHO Biologicals web site for the second round of public 2924

consultation from 1 February to 15 March 2016. 2925

2926

References 2927

1. Guidelines on clinical evaluation of vaccines: regulatory expectations. WHO Expert 2928

Committee on Biological Standardization: fifty-second report. Geneva: World Health 2929

Organization; 2004: Annex 1 (WHO Technical Report Series No. 924; 2930

http://www.who.int/biologicals/publications/trs/areas/vaccines/clinical_evaluation/035-2931

101.pdf, accessed 31 August 2015) 2932

2. WHO Technical Report Series website: 2933

http://who.int/biologicals/technical_report_series/en/. 2934

http://www.who.int/biologicals/publications/trs/areas/vaccines/clinical_evaluation/035-101.pdf

http://www.who.int/biologicals/publications/trs/areas/vaccines/clinical_evaluation/035-101.pdf

http://who.int/biologicals/technical_report_series/en/


3. Guidelines on nonclinical evaluation of vaccines. WHO Expert Committee on Biological 2935

Standardization: fifty-fourth report. Geneva: World Health Organization; 2005: Annex 1 2936

(WHO Technical Report Series, No. 927; 2937

(http://www.who.int/biologicals/publications/trs/areas/vaccines/nonclinical_evaluation/ANN2938

EX%201Nonclinical.P31-63.pdf?ua=1, accessed 31 August 2015). 2939

4. Guidelines for good clinical practice (GCP) for trials on pharmaceutical products. WHO 2940

Expert Committee on the Use of Essential Drugs: sixth report. Geneva: World Health 2941

Organization; 1995: Annex 3 (WHO Technical Report Series, No. 850; 2942

(http://whqlibdoc.who.int/trs/WHO_TRS_850.pdf?ua=1, accessed 31 August 2015). 2943

5. WHO good manufacturing practices for pharmaceutical products: main principles. WHO 2944

Expert Committee on Specifications for Pharmaceutical Preparations: forty-eighth report. 2945

Geneva: World Health Organization; 2014: Annex 2 (WHO Technical Report Series, No. 2946

986; 2947

http://www.who.int/entity/medicines/areas/quality_safety/quality_assurance/TRS986annex2.2948

pdf, accessed 31 August 2015). 2949

6. WHO GMP for Biological Products. WHO Expert Committee on Biological 2950

Standardization: XX report. Geneva: World Health Organization; 2015. Revision in process. 2951

(to be updated after ECBS 2015) 2952

7. Guidelines on nonclinical evaluation of vaccine adjuvants and adjuvanted vaccines. WHO 2953

Expert Committee on Biological Standardization: sixty-fourth report. Geneva: World Health 2954

Organization; 2013: Annex 2 (WHO Technical Report Series, 987; 2955

http://apps.who.int/iris/bitstream/10665/129494/1/TRS_987_eng.pdf?ua=1&ua=1, accessed 2956

31 August 2015) 2957

8. Guidelines on procedures and data requirements for changes to approved vaccines. WHO 2958

Expert Committee on Biological Standardization: sixty-fifth report. Geneva: World Health 2959

Organization; 2014: Annex 4 (WHO Technical Report Series, 993; 2960

http://www.who.int/biologicals/vaccines/Annex4_Guidelines_changes_to_approved_vaccin2961

es_eng.pdf?ua=1. 2962

9. Guidelines for independent lot release of vaccines by regulatory authorities. WHO Expert 2963

Committee on Biological Standardization: sixty-first report. Geneva: World Health 2964

http://www.who.int/biologicals/publications/trs/areas/vaccines/nonclinical_evaluation/ANNEX%201Nonclinical.P31-63.pdf?ua=1

http://www.who.int/biologicals/publications/trs/areas/vaccines/nonclinical_evaluation/ANNEX%201Nonclinical.P31-63.pdf?ua=1

http://whqlibdoc.who.int/trs/WHO_TRS_850.pdf?ua=1

http://www.who.int/entity/medicines/areas/quality_safety/quality_assurance/TRS986annex2.pdf

http://www.who.int/entity/medicines/areas/quality_safety/quality_assurance/TRS986annex2.pdf

http://www.who.int/biologicals/areas/vaccines/ADJUVANTS_Post_ECBS_edited_clean_Guidelines_NCE_Adjuvant_Final_17122013_WEB.pdf



http://apps.who.int/iris/bitstream/10665/129494/1/TRS_987_eng.pdf?ua=1&ua=1

http://www.who.int/biologicals/vaccines/Annex4_Guidelines_changes_to_approved_vaccines_eng.pdf?ua=1

http://www.who.int/biologicals/vaccines/Annex4_Guidelines_changes_to_approved_vaccines_eng.pdf?ua=1


Organization; 2013: Annex 2 (WHO Technical Report Series, No. 978; 2965

http://www.who.int/biologicals/TRS_978_Annex_2.pdf?ua=1, accessed 31 August 2015). 2966

10. Recommendations for the evaluation of animal cell cultures as substrates for the 2967

manufacture of biological medicinal products and for the characterization of cell banks. 2968

WHO Expert Committee on Biological Standardization: sixty-first report. Geneva: World 2969

Health Organization; 2013: Annex 3 (WHO Technical Report Series, No. 978, 2970

http://www.who.int/biologicals/vaccines/TRS_978_Annex_3.pdf?ua=1, accessed 31 August 2971

2015). 2972

11. Clinical Considerations for Evaluation of Vaccines for Prequalification. World Health 2973

Organization; 2010. 2974

http://www.who.int/immunization_standards/vaccine_quality/clinical_considerations_oct10.2975

pdf, accessed 31 August 2015. 2976

12. Immunization in practice: A practical guide for health staff. World Health Organization; 2977

2004. http://apps.who.int/iris/bitstream/10665/42991/1/9241546514.pdf, accessed 31 August 2978

2015. 2979

13. Expert consultation on the use of placebos in vaccine trials. World Health Organization; 2980

2013. http://apps.who.int/iris/bitstream/10665/94056/1/9789241506250_eng.pdf, accessed 2981

31 August 2015. 2982

14. Siegrist C-A, Vaccine Immunology” (Chapter 2) in Vaccines, ed. Plotkin SA, Orenstein 2983

WA, Offit PA, Sixth edition, Elsevier Saunders, 2012, pp. 14-32. 2984

15. Validation of Analytical Procedures: Text and Methodology. Q2(R1) document, 2985

International Conference on Harmonisation, 1996, pp. 6-13. 2986

16. Recommendations to Assure the Quality, Safety, and Efficacy of DT-based combined 2987

vaccines, Table 6.2, World Health Organization Expert Committee on Biological 2988

Standardization Technical Report Series, Geneva, Switzerland, vol. 980 annex 6, pp. 368-2989

70. 2990

17. Buttery JP, Riddell A, McVernon J, Chantler T, Lane L, Bowen-Morris J et al. 2991

Immunogenicity and safety of a combination pneumococcal-meningococcal vaccine in 2992

infants: a randomized controlled trial. JAMA 2005; 293 (14):1751-1758. 2993

18. Richmond P, Kaczmarski E, Borrow R, Findlow J, Clark S, McCann R et al. 2994

Meningococcal C polysaccharide vaccine induces immunologic hyporesponsiveness in 2995

http://www.who.int/biologicals/TRS_978_Annex_2.pdf?ua=1

http://www.who.int/biologicals/vaccines/TRS_978_Annex_3.pdf?ua=1

http://www.who.int/immunization_standards/vaccine_quality/clinical_considerations_oct10.pdf

http://www.who.int/immunization_standards/vaccine_quality/clinical_considerations_oct10.pdf

http://apps.who.int/iris/bitstream/10665/42991/1/9241546514.pdf

http://apps.who.int/iris/bitstream/10665/94056/1/9789241506250_eng.pdf


adults that is overcome by meningococcal C conjugate vaccine. J Infect Dis. 2000, 181 2996

(2):761-764. 2997

19. WHO consultation on influenza vaccines for pregnant and lactating women: Clinical data 2998

requirements for product labelling, 15 – 16 July 2014 Geneva, Switzerland. 2999

http://www.who.int/biologicals/vaccines/INFLUENZA_VACCINES_Executive_Summary_3000

FINAL_29_Sept_R.pdf, accessed 31 August 2015. 3001

20. Henao-Restrepo AM, et al. The ring vaccination trial: a novel cluster randomized controlled 3002

trial design to evaluate vaccine efficacy and effectiveness during outbreaks, with special 3003

reference to Ebola. BMJ 2015; 351;h3740. DOI: http://dx.doi.org/10.1136/bmj.h3740. 3004

21. Foppa IM, Haber M, Ferdinands JM, Shay DK. The case test-negative design for studies of 3005

the effectiveness of influenza vaccine. Vaccine 2013; 31: 3104-9. 3006

22. Global Manual on Surveillance of Adverse Events Following Immunization. World Health 3007

Organization 2014. DOI: 3008

http://www.who.int/vaccine_safety/publications/Global_Manual_on_Surveillance_of_AEFI.3009

pdf, accessed 31 August 2015. 3010

3011

http://www.who.int/biologicals/vaccines/INFLUENZA_VACCINES_Executive_Summary_FINAL_29_Sept_R.pdf

http://www.who.int/biologicals/vaccines/INFLUENZA_VACCINES_Executive_Summary_FINAL_29_Sept_R.pdf

http://dx.doi.org/10.1136/bmj.h3740

http://www.who.int/vaccine_safety/publications/Global_Manual_on_Surveillance_of_AEFI.pdf

http://www.who.int/vaccine_safety/publications/Global_Manual_on_Surveillance_of_AEFI.pdf


Appendix 1. Human Challenge Trials 3012

3013

There are many reasons a developer might wish to conduct with humans a “challenge-3014

protection” study that might normally be conducted in animals. Animal models are often quite 3015

imprecise in reflecting human disease and many infectious organisms against which a 3016

developer might wish to develop a vaccine are species-specific for humans. Human Challenge 3017

Trials may be safely and ethically performed in some cases, if properly designed and 3018

conducted. Tremendous insight into the mode-of-action and the potential for benefit in the 3019

relevant species, humans, may be gained from challenge trials. However, there are also 3020

limitations to what challenge trials may be able to ascertain, because like animal model 3021

challenge-protection studies, a human challenge trial represents a model system. Because there 3022

are often such significant limitations to animal models however, the model system of the 3023

human challenge trial may significantly advance, streamline, and/or accelerate vaccine 3024

development (1). 3025

3026

It will be important to consider the regulatory framework where the human challenge trial may 3027

be conducted, because in some countries, challenge stocks are expected to be handled in the 3028

same manner as vaccines and to be studied under a Clinical Trial Authorization (Approval, 3029

CTA), whether or not an investigational vaccine is to be used in the same clinical investigation 3030

protocol. For example, a challenge trial might be conducted to titrate the challenge organism in 3031

humans before using the challenge in a vaccine study, in order to know the proper dose of the 3032

challenge organism to give and to characterize the symptoms, kinetics, shedding, 3033

transmissibility, and so forth to expect from the challenge. In such cases (when challenge 3034

should be studied under CTA), there is greater clarity about regulatory expectations, including 3035

quality of the challenge stock to be used, as the CTA regulations or requirements would apply. 3036

However, in many countries, because the challenge stock is not itself a medicinal product, such 3037

studies would not be under the purview of the NRA’s review and approval and much less 3038

clarity exists on regulatory expectations and quality matters in such cases. Ideally, a challenge 3039

stock should match in quality terms what is expected of an investigational vaccine at the same 3040

clinical Phase of development (understanding that a pathogenic challenge strain will not have 3041

the “safety” of a hopefully innocuous vaccine). Likewise, ideally a human challenge study 3042


should match the same expectations for conduct of a vaccine study, e.g., compliance with 3043

GCP, approval of a CTA. However, there may not exist a regulatory framework to promulgate 3044

such expectations in the country where the challenge study is to be conducted. Thus, it may be 3045

necessary for regulators to consider and develop an appropriate regulatory pathway or 3046

framework for the quality of the challenge stock and the conduct of the challenge study, when 3047

clarity is not apparent in their existing system. This may require new legislation to give 3048

regulators the necessary authority, and it is encouraged that regulators should have this 3049

authority. Trial sponsors, vaccine developers, researchers, and so on should determine from the 3050

relevant NRA what regulatory expectations they may have when clarity does not exist, if the 3051

human challenge study is intended to support the development of a vaccine candidate they 3052

would like to ultimately license (i.e. gain marketing authorization). 3053

3054

It is also important to note that not all diseases for which vaccines might be developed are 3055

suitable to consider conducting human challenge trials. In many cases, human challenge with a 3056

virulent or even a potentially attenuated organism would not be considered ethical or safe. For 3057

example, if an organism causes a high case fatality rate (or there is a long and uncertain latency 3058

period) and there are no existing therapies to prevent or ameliorate disease and preclude death, 3059

then it would not be appropriate to consider human challenge trials with such an organism. 3060

However, when the disease an organism causes has an acute onset and can be readily and 3061

objectively detected and existing efficacious treatments (whether curative or palliative) can be 3062

administered at an appropriate juncture in disease development to prevent significant 3063

morbidity (and eliminate mortality), a human challenge trial might be considered. 3064

3065

1. Purposes of human challenge trials 3066

A developer may conduct human challenge trials to accomplish one or more of a number of 3067

aims. The aims of the study determine what clinical Phase the study may be considered to be. 3068

Human challenge trials are often a type of efficacy study, but not all would be considered a 3069

“Phase 3” study. Purposes of human challenge trials could include one or more of the 3070

following: 3071

Characterization of the challenge stock and model system: titration, symptoms, kinetics, 3072

shedding, transmissibility, etc. 3073


Clearer understanding of pathogenesis of and immunity to the organism in order to guide 3074

decisions on what (type and/or quantity) immune responses a vaccine might need to 3075

accomplish in order to protect against that disease, i.e. insight for vaccine design (studies for 3076

this purpose may be referred to as experimental medicine studies) 3077

Identification of potential immune correlates of protection (ICP, which would then require 3078

validation in a traditional efficacy study) 3079

Identification of optimal trial design for Phase 3 traditional efficacy trial(s), e.g. case 3080

definitions, endpoints, study design aspects 3081

Generation of appropriate hypotheses to be formally tested in traditional efficacy trials 3082

Proof-of-concept that a particular vaccine candidate might be capable of protection or not 3083

Down- or Up-selection among various potential lead vaccine candidates to advance only the 3084

best to large Phase 2b or Phase 3 efficacy trials and to eliminate those that are unworthy of 3085

advancement 3086

De-risk or “left-shift”1 risk of failure in a vaccine development program 3087

Comparison of vaccine performance in endemic settings vs. in efficacy trial population2, 3088

including evaluating impact of prior immunity 3089

Support emergency use of an investigational vaccine, e.g. in a pandemic 3090

Basis for licensure (this purpose would generally be an exception rather than the rule) 3091

Exploration post-licensure whether immunity to vaccination wanes and if or when booster 3092

doses might be required for durable protection3 3093

Others 3094

Not all situations would support accomplishing each of the aims above. For example, if the 3095

human challenge model system does not adequately mimic the wild-type disease and situation 3096

in which a vaccine would need to protect, then a human challenge trial would not be usable as 3097

a basis for licensure. But, it might still serve well one or more of the other purposes above. It 3098

1 When looking at a timeline of vaccine development graphed from early to the left and late to the right, shifting the

risk of failure earlier in the timeline, or left, could result in significant cost (and resource)-savings and minimize lost

opportunity costs by abandoning an unpromising candidate before taking greater expenditures from higher phase

clinical trials, not to mention minimizing risk to human subjects by not conducting large efficacy studies of vaccines

that would not prove efficacious 2 Target population in a particular country may have a higher rate of individuals with e.g., sickle cell trait or

different nutritional status or greater parasitic load in “normal” flora, any of which might affect immune

responsiveness and thus, efficacy, compared to the efficacy trial population 3 This might entail challenge study in adults to extrapolate when children might need booster doses


might even be considered by regulators as supportive of licensure, but not a sole or primary 3099

basis. 3100

3101

2. Purpose influences study design, which influences regulatory use and decision-making 3102

Obviously, the aim of the human challenge trial guides its study design. Consequently, even 3103

for the same disease, the challenge model may vary depending on the purposes and design of 3104

the study to be conducted. In some cases (e.g. to serve as a basis for licensure or to identify 3105

appropriate efficacy trial design and case definitions), the challenge model might need to 3106

mimic as closely as feasible wild-type disease. In other cases, consideration might be given to 3107

use of an attenuated challenge organism (e.g., an earlier but under-attenuated vaccine 3108

candidate) or a model system in which objective early signs (e.g. parasitaemia, viraemia) 3109

signaling onset of disease symptoms, which could trigger initiation of treatment to prevent 3110

actual disease onset or morbidity. 3111

3112

Another important consideration for a human challenge model system would be its positive 3113

and negative predictive utility. If used for down-selection or de-risking, the negative predictive 3114

utility of the model to identify vaccine candidates that would not warrant advancement into 3115

large human efficacy studies should be high. If intended to be used for licensure, the positive 3116

predictive utility of the model system would need to be nearly as compelling and credible as a 3117

traditional efficacy trial might be. Thus, the purpose of the study would influence the design, 3118

which would in turn influence the conclusions about and the decisions that might be made 3119

from the study results. 3120

3121

3. Some key ethical considerations 3122

Ethics in clinical trials, as in medicine, follow the precept of “do no harm.” By their nature 3123

(intentionally infecting humans with disease-causing organisms), human challenge trials would 3124

seem to fly in the face of this basic precept. Further, clinical trials should be designed and 3125

conducted in a manner that minimizes risks to human subjects while maximizing the potential 3126

to benefit. Consideration must be given both to potential individual risks and benefits, as well 3127

as to potential societal benefits (and risks, such as release into the environment of a pathogen 3128

that might not otherwise be present). Provisions in clinical trial ethics are made for situations 3129


in which there may be greater than minimal risk but no (or little) potential for individual 3130

benefit, but when knowledge may be gained to the benefit of the larger societal population 3131

with whom the potential trial participant shares significant characteristics. Justification for 3132

asking trial participants to accept the risk from a challenge may take some considerations from 3133

the justifications that support inclusion of placebos in controlled clinical trials. 3134

3135

Acknowledgement is due to the reality that some individuals are greater risk-takers than others, 3136

while some individuals are quite risk-averse and would not be accepting of the risk of 3137

receiving a challenge. Key to asking individuals to accept the risk from a challenge study in 3138

which they may not except to receive individual benefit is the element of informed consent. 3139

Adults may consent when they are well-informed and understand what risks they are accepting 3140

to take, even if those risks may be considerably greater than minimal (e.g. accepting that they 3141

will develop an acute, but manageable, disease that will resolve but in the meantime may cause 3142

considerable morbidity, e.g. severe diarrhea managed with fluid and electrolyte replacement). 3143

Thus, in appropriate situations, it can be considered ethical to ask informed adults to consent to 3144

volunteer and participate in a human challenge trial whether they will receive an 3145

investigational vaccine that may or may not protect them from the challenge organism, a 3146

placebo that will not protect them, or only the challenge organism itself. However, accepting 3147

such risks requires absolutely the elements of voluntary consent based on truly being informed. 3148

It is for this reason (need for truly informed consent), consideration of conducting human 3149

challenge studies in children or any other vulnerable population, who would have diminished 3150

capacity to give informed consent, would not be deemed acceptable at this time. 3151

3152

The need to minimize risks to subjects in clinical trials calls for due consideration to whether 3153

or not the challenge organism need be pathogenic or not, or to what degree. As stated above, 3154

the aim or purpose of the study may drive this decision, but the ethics of minimizing to the 3155

extent feasible within the frame of sound science any risks to human subjects should also bear 3156

due consideration in this regard. It should also be obvious that the credibility of the data to 3157

support regulatory decision-making need be taken into account. 3158

3159

References 3160


3161

1. Sheets RL, Fritzell B, Aguado de Ros MT, Human Vaccine Challenge Trials in Vaccine 3162

Development. Biologicals, submitted. 3163

3164

3165

Guidelines on Clinical Evaluation of Vaccines: Regulatory ... · 7 Guidelines on Clinical...

Documents

Transcript of Guidelines on Clinical Evaluation of Vaccines: Regulatory ... · 7 Guidelines on Clinical...